muhammad et al. aspect-based sentiment analysis on amazon product reviews | 94 aspect-based sentiment analysis on amazon product reviews muhammad abubakar*, amir shahzad, husna abbasi comsats university islamabad abbottabad campus pakistan, pakistan. *corresponding email: abubakarhameedch@gmail.com a b s t r a c t s a r t i c l e i n f o the focus of this paper was on amazon product reviews. the goal of this is to study is two (nlp) for evaluating amazon product review sentiment analysis. customers can learn about a product's quality by reading reviews. several product review characteristics, such as quality, time of evaluation, material in terms of product lifespan and excellent client feedback from the past, will have an impact on product rankings. manual interventions are required to analyse these reviews, which are not only time consuming but also prone to errors. as a result, automatic models and procedures are required to effectively manage product reviews. (nlp) is the most practical method for training a neural network in this era of artificial intelligence. first, the naive bayes classifier was used to analyse the sentiment of consumer in this study. the (svm) has categorized user sentiments into binary categories. the goal of the approach is to forecast some of the most important characteristics of an amazon-based product reviews, and then analyse customer attitudes about these aspects. the suggested model is validated using a largescale real-world dataset gathered specifically for this purpose. the dataset is made up of thousands of manually annotated product reviews gathered from amazon. after passing the input via the network model, (tf) and (idf) pre-processing methods were used to evaluate the feature. the outcomes precision, recall and f1 score are very promising. article history: received 18 dec 2021 revised 20 dec 2021 accepted 25 dec 2021 available online 26 dec 2021 aug 2018 __________________ keywords: naïve bayes, text classification algorithms , natural language processing, support vector machines, nlp, svm international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(2) (2021) 94-99 95 | international journal of informatics information system and computer engineering 2(2) (2021) 94-99 1. introduction amazon is the largest online retailer in the world, as well as a significant cloud computing service provider (rain, 2013). the company began as a book seller but has now evolved to include a wide range of consumer items and digital media, including the kindle e-reader, kindle fire tablet, and fire tv., a streaming media adaptor are among the company's own electronic devices. people nowadays prefer to trade things on an e-commerce website rather than at a physical store because of the time savings and convenience (bhat et al., 2015). before purchasing a product, it is usual practice to read the product review. the consumer's opinion of the product has been swayed either positive or negative by the reviews. thousands of reviews were read, on the other hand, is an unnatural feat. in this era of everimproving natural language processing algorithms, it takes time to wade through hundreds of comments to identify a product that uses a polarized review of a specific category to assess its popularity among consumers all around the world. this project aims to categorize customers' positive and negative product reviews, as well as construct a supervised learning model to polarize a wide range of reviews. according to an amazon study from last year, 88% customers from internet trust reviews as much as a personal suggestion (dey et al., 2020). with a powerful remark, the credibility of an internet product with a high number of positive reviews is established. the absence of reviews, books, or any other thing on the internet creates a sense of distrust among potential customers. pre-processing is used in this study to minimize the multidomain sentiment dataset's dimensionality of the features applied. following that, any frequent words above a certain threshold value are considered characteristics (haque et al., 2018). 2. related work this section presents the results of the classic schema polarization analysis based on user reviews on the amazon ecommerce website (xiao et al., 2021). the criteria for compositional sentiment were set by zhang et al. to find out how much textual sentiment there is. the system makes clear use of machine learning. in this work, film reviews were classified into binary classes using (svm) and naive bayes classifiers (joseph, 2020). the accuracy of the naive bayes model has been improved, while the svm model has been extended. to summarize, there have been no studies comparing (svm) with the naive bayes classifier. a comparison of two approaches (nlp) to analyze amazon product evaluation sentiment is presented in this study (more et al., 2020). comparative polarity analysis on amazon product reviews using algorithms has also been carried out to evaluate the sentiment of amazon product reviews (karthikayini et al., 2017). in his research, dadhich uses a rule-based hybrid to be able to create an automatic comment analyzer (dadhich et al., 2022). salmony also conducted a survey on amazon product reviews to assist in customer decision making (salmony et al., 2021). 3. methodology amazon, as seen by the numerous evaluations accessible, is one of the most well-known e-commerce companies. the dataset was unlabeled, thus it needed to muhammad et al. aspect-based sentiment analysis on amazon product reviews | 96 be labelled before it could be used in a supervised learning model (pandey, 2019). only amazon product feedback, specifically book feedback, was used for this study activity. to evaluating polarization, about 1, 47,000 book evaluations were analyzed. data collecting was completed as the first step in the data labelling process. manual labelling is impractical for a human to do because the dataset contains a high number of reviews. the term (tf) and (idf), elimination of relevant nouns and frequent noun identifier methods were used to extract the dataset's features (jagdale et al., 2019). tf-idf: tf-idf is a retrieval strategy that considers the frequency of a phrase (tf) as well as the (idf). tf and idf scores are assigned to each word or phrase. the tf and idf product results of a term, on the other hand, refer to the tf-idf weight of that term. as a result, the tf of a word represents its frequency, whereas the idf is a metric for what percentage of the corpus is occupied by a term. the content will always be among the top search results if words have a high tf-idf content weight, allowing anyone to avoid stop words while also effectively locating words with a higher search volume but a lower level of competition (fang, 2015). 4. results and discussion the purpose of this part is to assess the experiment's performance. evaluating metrics are important in determining classification efficiency, and assessing accuracy is the easiest way to do so. the system is assessed using three widely used statistical measures: the f-measure, which is generated from a confusion matrix, is derived from recall, precision, and the f-measure. the confusion matrix divided into four categories true positive, true negative, false positive, and false negative (see figures 1 and 2). true positive describes a situation in which the system accurately anticipates the positive class. false-positive highlights a situation in which the scheme predicts the positive class inaccurately. tabulator form is used to show the (svm) confusion matrix and the naive bayes classifier a separate tabular format is used to display both the statistical measurement and the npl (table 1). table 1. svm confusion matrix positive 3694 neutral 158 negative 90 in the train dataset, we have 3694 (~95.1%) sentiments labelled as positive, and 158 (~4%) sentiments labelled as neutral and 90(~2.35%) sentiments as negative. so, it is an imbalanced classification problem. naive bayes [[0 0 24] [0 0 39] [0 0 937]] precision recall f1-score support 0 0.00 0.00 0.00 24 1 0.00 0.00 0.00 39 2 0.94 1.00 0.97 937 micro avg 0.94 0.94 0.94 1000 macro avg 0.31 0.33 0.32 1000 97 | international journal of informatics information system and computer engineering 2(2) (2021) 94-99 weighted avg 0.88 0.94 0.91 1000 accuracy: 93.7 precision refers to the ratio of predicted positive cases to total positive instances indicated by the equation. tf/idf vectorizer and logistic regression for under sampled data [[10 6 8] [15 7 17] [314 195 428]] precision recall f1-score support 0 0.03 0.42 0.06 24 1 0.03 0.18 0.06 39 2 0.94 0.46 0.62 937 micro avg 0.45 0.45 0.45 1000 macro avg 0.34 0.35 0.24 1000 weighted avg 0.89 0.45 0.58 1000 accuracy: 44.5 characteristic of logistic regression of under sampled data figure 1. true and false positive rate under sampled data tf/idf and logistic regression for over sampled data [[13 3 8] [10 10 19] [214 171 552]] precision recall f1-score support 0 0.05 0.54 0.10 24 1 0.05 0.26 0.09 39 2 0.95 0.59 0.73 937 micro avg 0.57 0.57 0.57 1000 macro avg 0.35 0.46 0.31 1000 weighted avg 0.90 0.57 0.69 1000 accuracy: 57.49999999999999 logistic regression on over-sampled data is performing better than undersampled data. muhammad et al. aspect-based sentiment analysis on amazon product reviews | 98 characteristic of logistic regression of over sampled data figure 2. true and false over sampled data neural network [[9 2 13] [0 12 27] [2 8 927]] precision recall f1-score support 0 0.82 0.38 0.51 24 1 0.55 0.31 0.39 39 2 0.96 0.99 0.97 937 micro avg 0.95 0.95 0.95 1000 macro avg 0.77 0.56 0.63 1000 weighted avg 0.94 0.95 0.94 1000 using class-weights does not improve the performance. 3. conclusion in order to investigate the polarisation of amazon product ratings, this study was able to compare svm and naive bayes classifiers. following the preprocessing step, almost 2250 features and over 6000 datasets were used to train the models. the svm classifier in this system has a precision of 0.00 percent, a recall of 0.00 percent, f1 score 0.00 percent. the model yields svm and naive bayes with 93.7 percent accuracy, respectively, which is confirmed to be superior to traditional approaches. with a higher accuracy rate, the (svm) can polarise amazon product feedback, according to the findings of experiments. 99 | international journal of informatics information system and computer engineering 2(2) (2021) 94-99 references bhatt, a., patel, a., chheda, h., & gawande, k. (2015). amazon review classification and sentiment analysis. international journal of computer science and information technologies, 6(6), 5107-5110. dadhich, a., & thankachan, b. (2022). sentiment analysis of amazon product reviews using hybrid rule-based approach. in smart systems: innovations in computing (pp. 173-193). springer, singapore. dey, s., wasif, s., tonmoy, d. s., sultana, s., sarkar, j., & dey, m. (2020, february). a comparative study of support vector machine and naive bayes classifier for sentiment analysis on amazon product reviews. in 2020 international conference on contemporary computing and applications (ic3a) (pp. 217-220). ieee. fang, x., & zhan, j. (2015). sentiment analysis using product review data. journal of big data, 2(1), 1-14. haque, t. u., saber, n. n., & shah, f. m. (2018, may). sentiment analysis on large scale amazon product reviews. in 2018 ieee international conference on innovative research and development (icird) (pp. 1-6). ieee. jagdale, r. s., shirsat, v. s., & deshmukh, s. n. (2019). sentiment analysis on product reviews using machine learning techniques. in cognitive informatics and soft computing (pp. 639-647). springer, singapore. joseph, r. p. s. (2020). amazon reviews sentiment analysis: a reinforcement learning approach (doctoral dissertation, ms thesis, griffith college dublin, ireland). karthikayini, t., & srinath, n. k. (2017, december). comparative polarity analysis on amazon product reviews using existing machine learning algorithms. in 2017 2nd international conference on computational systems and information technology for sustainable solution (csitss) (pp. 1-6). ieee. more, g., behara, h., & suresha, a. m. (2020). sentiment analysis on amazon product reviews with stacked neural networks. no. october. pandey, p., & soni, n. (2019, february). sentiment analysis on customer feedback data: amazon product reviews. in 2019 international conference on machine learning, big data, cloud and parallel computing (comitcon) (pp. 320-322). ieee. rain, c. (2013). sentiment analysis in amazon reviews using probabilistic machine learning. swarthmore college. salmony, m. y. a., & faridi, a. r. (2021, april). supervised sentiment analysis on amazon product reviews: a survey. in 2021 2nd international conference on intelligent engineering and management (iciem) (pp. 132-138). ieee. xiao, y., qi, c., & leng, h. (2021, march). sentiment analysis of amazon product reviews based on nlp. in 2021 4th international conference on advanced electronic materials, computers and software engineering (aemcse) (pp. 12181221). ieee. 71 | international journal of informatics information system and computer engineering 3(1) (2022) 71-79 bts application: online thesis consultation bella hardiyana school of information science japan advanced institute of science and technology, japan *corresponding email: bella.hardiayana@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o the learning process at universities is hindered by the covid-19 pandemic, so activities that should be carried out face-to-face must be done online. one of the activities that are hampered is thesis consultation. thesis consultation should be done directly, face to face, and verified on the attendance card, and it cannot be carried out as usual. the consultation can only be done online by sending files to the supervisor and then reviewing the results of the work. however, by doing it online, it will not be easy to fill out the attendance card that must be signed in person. the signing process becomes online, by sending the digital version of attendance card to the supervisor, then be signed by the supervisor and sent again to the student. the purpose of this research is to design a thesis consultation information system in which everything is centralized and documented in one platform. this research used qualitative descriptive analysis method and system development method using prototype. the results showed that the system design that was built could help become a medium for exchanging files between supervisors and students. the output of this system will be documented. the history of each counsel carried out will be recorded so that the attendance card will be filled automatically. article history: received 25 may 2022 revised 30 may 2022 accepted 10 june 2022 available online 26 june 2022 aug 2018 __________________ keywords: technology, information system, computer science, application, assignment, thesis international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 71-79 bella hardiyana. bts application: online thesis assignment guidance | 72 1. introduction the corona virus is the cause of the covid-19 pandemic which is currently spreading in various countries (skovlund et al., 2021). this virus is an infectious disease virus that has a very fast spread with human fluids as the medium of transmission (singhal, 2020). this virus target people indiscriminately and tends to be more dangerous if it affects the elderly and people who have a history of previous severe illness (etard et al., 2020). with the outbreak of covid-19, many activities have been hindered, if not stopped. covid-19 has made a country's economy paralyzed, interrupted the distribution of logistics, and hindered the education sector is because the spread of this virus spreads through physical and liquid contact so that it spreads very quickly (alagu et al., 2021). this certainly has a lot of impact on the sectors of daily life, one of which is the education sector. offline learning activities must change to online to avoid the spread of the virus so that all students are forced to adapt to the new method. in addition, the university's higher education sector is also constrained, for example in the process of implementing the final project. the final project also has a consultation process which is usually carried out directly by the lecturers and their student group, but it is now constrained by the policy of limiting social activities so that all must be connected through the online system (kintama et al., 2021). therefore, to solve these problems, it is necessary to build media to be a liaison between the consultation process between lecturers and students who can monitor the progress of the final project. in china, the use of online learning platforms was practiced even before the covid-19 pandemic. however, during the pandemic, the use of online platforms has skyrocketed, but it is not uncommon for people to worry about the security and speed of video conference data transfer. for this reason, video conferencing service providers move quickly to overcome these problems by updating to minimize application bugs (han et al., 2021). in addition, studies conducted in japan indicate that many universities are not ready organizationally or operationally when facing a pandemic. therefore, many universities collaborate with other universities to establish cooperation in the implementation of online classes to support their learning. in addition, many universities are preparing post-pandemic scenarios so that the learning carried out remains relevant to the surrounding community (izumi et al., 2021). in line with this, online education according to a survey of several students in india stated that this method was a feasible alternative during this pandemic. although the survey conducted stated that 65.9% felt that learning through physical classrooms was more effective than through online. for this reason, students hope to optimize this online learning by delivering more diverse materials such as case studies, gamification, and interactive classes (chakraborty et al., 2021). based on previous research, this study raised the theme of online learning during the pandemic, especially in the final project guidance process that was affected by being online with the covid-19 pandemic. the purpose of this research is to design a thesis consultation information system in which everything is centralized and documented in one platform. this 73 | international journal of informatics information system and computer engineering 3(1) (2022) 71-79 research used qualitative descriptive analysis method and system development method using prototype. the results show that the system design that was built can help become a medium for exchanging files between supervisors and students. 2. method this research method used descriptive qualitative analysis with object-oriented systems approach method. the concept of object-oriented approach makes developers focus on creating classes which are the blueprints of an object system. this concept can divide the system components into several objects that interact with each other to run the system (aman, 2021). while in data collection using observation techniques and direct interviews on the object of research. interviews are used to determine the needs of users who will use this application and observations are made on activities that run before the system (greer et al., 2020). in the development of this system using the rapid application development (rad) method. rapid application development (rad) is a software development process model that moves linearly over and over in development but is limited by a short time because this method is specifically for systems that are not too complex (pricillia, 2021). because the stages used will work a lot at one stage of development before the final stage of implementation. the following are the stages of the rad development method (rosmalia et al., 2021) (fig. 1). fig. 1. rapid application development (rad) method 3. results and discussion the development of this thesis consultation information system used the php programming language with the codeigniter framework and from the database supported by the mysql dbms. in the proposed system, this system has 2 main functions that are used by 2 users or users, namely lecturers and students. each user has a similar function in the system, but both have their own rights and characteristics. student users can consult with their supervisor, by submitting the files through the menu provided (upload guidance), after that, the student waits for the lecturer to verify his attendance and after that students can see the comments in the menu provided (revision consultation). students can also see the history of consultation that he did together with his supervisor (consultation history). the following is a design use case diagram of the thesis consultation information system (fig. 2). bella hardiyana. bts application: online thesis assignment guidance | 74 fig. 2. use case thesis consultation diagram in this information system, students can upload a draft of the file-to-beconsulted in stages, for example by uploading it according to the chapter they consulted. in addition to the facility for uploading thesis drafts, there are also facilities from lecturers to provide comments in which these comments will be recorded on the attendance card. each stage of guidance carried out by the student if it is in accordance with the results of the revision, the lecturer will provide validation through the system which will be recorded on the attendance card in the form of initials based on the stages of consultation carried out. in addition to thesis draft consultation, it is also possible to test programs that have been made by students. students can test their program by recording their screen then upload it on youtube and attach the link to the program testing consultation form on this information system. as with the previous stages in the program testing consultation, lecturers are also given the feature to provide comments on what has been presented by students. the lecturers are also given an approval feature to validates student submission at this stage. after all the stages of consultation are carried out and validated by the supervisor, students only need to download the attendance card file to be used as evidence for conducting consultation as a condition for the thesis trial. in designing information systems, activity diagrams serve as an overview of the system flow and what the system can do. the following is an illustration of the designed activity diagram (fig. 3). 75 | international journal of informatics information system and computer engineering 3(1) (2022) 71-79 fig. 3. thesis consultation activity diagram bella hardiyana. bts application: online thesis assignment guidance | 76 the thesis consultation process starts from students accessing the page then uploading the results of their thesis work. the form contains a brief description of the draft uploaded in this tutorial. then upload the draft thesis file to be reviewed by the lecturer concerned. in this form, students can choose which chapter will be the topic of the consultation so that the lecturer can monitor the progress of the student group under his guidance. fig. 4 is the interface for the thesis draft upload page. fig. 4. upload thesis draft by student after the guidance draft has been uploaded successfully, it will be entered on the supervisor page. the supervisor will download the previously uploaded file to be able to do a review. after the review is done, the lecturer will make a response to the student in the form of a note of improvement that needs to be done by the student. in this form, lecturers can also upload their review files to be re-sent to students so that students can see in detail the parts that need to be improved. at this verification stage, if you feel that the draft is appropriate, you can change the status of the chapter's guidance to "accepted". the guidance verification page can be seen in fig. 5 and 6. 77 | international journal of informatics information system and computer engineering 3(1) (2022) 71-79 fig. 5. review draft skripsi by lecturer fig. 6. consultation history by student lecturers can check the student attendance list through the consultation list menu. the lecturer can also revise files submitted by the students in the consultation verification menu can see the consultation history with students in the consultation history menu. in fig. 7 is the design of the consultation history interface that can be accessed by lecturers. bella hardiyana. bts application: online thesis assignment guidance | 78 fig. 7. list of students in guidance 4. conclusion the making of this thesis consultation information system require real data and problems that actually occur in the educational environment, namely the educational environment of the universitas komputer indonesia (unikom). with the covid-19 pandemic making it difficult for students to conduct thesis guidance directly, an information system was created with the php programming language and using the mysql database. with the creation of this information system, it will make it easier for students to carry out the thesis guidance procession because students do not need to come directly to campus for guidance, but only by opening the platform provided students can do thesis guidance, in the preparation of this information system, of course, the author hopes this system can make it easier for lecturers and students to carry out the educational process in this pandemic situation. references skovlund, c. w., friis, s., dehlendorff, c., nilbert, m. c., & mørch, l. s. (2021). hidden morbidities: drop in cancer diagnoses during the covid-19 pandemic in denmark. acta oncologica, 60(1), 20-23. singhal, t. (2020). a review of coronavirus disease-2019 (covid-19). the indian journal of pediatrics, 87(4), 281-286. 79 | international journal of informatics information system and computer engineering 3(1) (2022) 71-79 etard, j. f., vanhems, p., atlani-duault, l., & ecochard, r. (2020). potential lethal outbreak of coronavirus disease (covid-19) among the elderly in retirement homes and long-term facilities, france, march 2020. eurosurveillance, 25(15), 2000448. alagu lakshmi, s., shafreen, r. m. b., priya, a., & shunmugiah, k. p. (2021). ethnomedicines of indian origin for combating covid-19 infection by hampering the viral replication: using structure-based drug discovery approach. journal of biomolecular structure and dynamics, 39(13), 4594-4609. kintama, a. y., larasati, d. a., & yuliana, l. (2021). bimbingan skripsi daring selama pademi covid-19 pada mahasiswa pgsd uwks: hambatan dan solusi. trapsila: jurnal pendidikan dasar, 3(1), 57-71. han, x., zhou, q., shi, w., & yang, s. (2021). online learning in vocational education of china during covid-19: achievements, challenges, and future developments. journal of educational technology development and exchange (jetde), 13(2), 4-31. izumi, t., sukhwani, v., surjan, a. and shaw, r. (2021), "managing and responding to pandemics in higher educational institutions: initial learning from covid19", international journal of disaster resilience in the built environment, vol. 12 no. 1, pp. 51-66. chakraborty, p., mittal, p., gupta, m. s., yadav, s., & arora, a. (2021). opinion of students on online education during the covid‐19 pandemic. human behavior and emerging technologies, 3(3), 357-365. aman, m. (2021). pengembangan sistem informasi wedding organizer menggunakan pendekatan sistem berorientasi objek pada cv pesta. jurnal janitra informatika dan sistem informasi, 1(1), 47-60. greer, b. d., mitteer, d. r., briggs, a. m., fisher, w. w., & sodawasser, a. j. (2020). comparisons of standardized and interview‐informed synthesized reinforcement contingencies relative to functional analysis. journal of applied behavior analysis, 53(1), 82-101. pricillia, t. (2021). perbandingan metode pengembangan perangkat lunak (waterfall, prototype, rad). jurnal bangkit indonesia, 10(1), 6-12. rosmalia, l., jaroji, j., & teddyyana, a. (2021). aplikasi pendataan dan monitoring industri kecil dan menengah (ikm) menggunakan metode rapid application development. zonasi: jurnal sistem informasi, 3(2), 71-86. faez m. hassan, hussein abdelwahab mossa. image mosaicking using low-distance...| 44 image mosaicking using low-distance high-resolution images captured by an unmanned aerial vehicle faez m. hassan, hussein abdelwahab mossa physics department, college of education, mustansiriyah university, baghdad, iraq a b s t r a c t s a r t i c l e i n f o regional surveys will have a high demand for coverage. to adequately cover a large area while retaining high resolution, mosaics of the area from a variety of scenes can be created. this paper describes a mosaicking procedure that consists of a series of processing steps used to combine multiple aerial images. these images were taken from cropcam unmanned aerial platform flight missions over the desired area to quickly map a large geographical region. the results of periodic processing can be compared and analyzed to monitor a large area for future research or during an emergency situation in the covered area. digital imagery captured from the air has proven to be a valuable resource for studying land cover and land use. for this study, airborne digital camera images were chosen because they provide data with a higher spatial resolution for trying to map a small research area. on board the uav autopilot, images were captured from an elevation of 320 meters using a standard digital camera. when compared to other airborne studies, this technique was less expensive and more cost effective. according to this study, onboard a uav autopilot, a digital camera serves as a sensor, which can be helpful in planning and developing a limited coverage area after mosaicking. article history: received 18 nov 2021 revised 20 nov 2021 accepted 25 nov 2021 available online 26 dec 2021 __________________ keywords: image mosaic, crop cam uav, aerial photography international journal of informatics information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 2(2) (2021) 44-52 45 | international journal of informatics information system and computer engineering 2(2) (2021) 44-52 1. introduction aerial photography serves as a common foundation for large-scale mapping. it has been widely used for creating and updating maps, as well as for keeping gis databases up to date (neteler, m., & mitasova, h. 2004; xu, y et al., 2016). in remote sensing and geoformation sciences, the term "mosaicking" is frequently used when two or more contiguous images are stitched together to create a single image file (aber, et al., 2010). stitching functions are now available in the majority of digital camera software, and there is a wide range of both free and commercial panorama software available for fully automating the merging of combining multiple photographs into larger composites. these tools have the potential to produce visually appealing results, but they are not geometrically correct, which means they can still detect the image tiles with an angular skew and straight edges (aber, et al., 2010). one of the most significant image data processing techniques in uav systems is mosaic in real time, which allows the uav images that have been georeferenced to be combined with geographic information for quick reaction to time-sensitive events (zhou, et al., 2006; kim, et al., 2017). aerial photographs can be used to study changes in the earth's features as time passes. those images are especially useful in analyses of land cover because they compare older data sets with new data sets, which can be available for a wide range of studies (ren, et al., 2017). information on current land use allows agencies and researchers to identify patterns in land cover and, as a result, make more informed decisions about analyses of development suitability, proposed land uses, and long-term planning (gómez-candón, et al., 2014). the data can show how development has changed over time, which can be applied as a guide for future research on land cover (ahmad, a., 2011; hassan, et al., 2010; gomarascs, m. a. 2009). the digital images captured are available in a short time and are accompanied by latitude, longitude, and altitude coordinates (zhao., et al., 2019). by manipulating the visualization of digital images, the user can keep track of what is going on at the ground, observe the most recent developments, and prevent problems from spiraling out of control. the unmanned aerial vehicle (uav) can be hand-launched and can autonomously fly from takeoff to landing (cropcam, 2008; felderhof, et al., 2008). both flights were made to capture visible imagery with a resolution of ground level of 9 cm over the selected area in each flight, and all of the images obtained were in jpeg format. the visible images that resulted demonstrated a clear distinction between urban and green land surfaces (avola, et al., 2018). flights of uav’s have been completed successfully at all of the research sites selected for this study. the number of photos taken during each flight plan was adequate for covering the research area. a mosaic image is a fabricated composition created from a series of images obtained by comprehending the geometric relationships between images (fuyi, et al., 2012; lim, et al., 2009; hassan, et al., 2011). the entire survey area should be covered by aerial images with a sufficient amount of overlap between them. typically, the degree of overlap in route direction should be between 60% and 65% with no less than 53%, while in the lateral direction, it should be between 30% and 40% with no less than 15% (wang, et al., 2007). each fly file was set up in this study with a 60% faez m. hassan, hussein abdelwahab mossa. image mosaicking using low-distance...| 46 overlap along the flight end lap runs and a 30% overlap between the side lap runs. the images that were taken in this study had a lot of overlap between them, both inside and outside of the runs. this meant that the images could be stitched together into a good mosaic for further analysis. 2. method penang island, which is located in northern malaysia between latitudes 5o 12' n and 5o 30' n and longitudes 100o 09' e and 100o 26' e, was chosen as a study area. the cropcam uav system was used to collect images for this study, as shown in figure 1. the aircraft was outfitted with navigation and autopilot systems that allowed it to follow predetermined waypoints and thus acquire the target area. furthermore, to collect digital remote sensing images of the study area, a pentax optio a40 digital camera was employed in the form of a low-cost imaging sensor system affixed to the body of the uav. pentax's first digital camera with a resolution of 12.0 effective megapixels, the optio a40 is capable of producing images with extreme precision and high resolution. fig. 1. cropcam uav system used in this study. flights were conducted for each study site to capture visible imagery in order to cover the entire study site, with high ground resolution. in this study, the imagery was captured by the uav at a low altitude (320 meters) above ground level, allowing imagery to be obtained even when there was cloud cover, giving it a competitive advantage over manned aircraft and satellite imagery. the software packages autopano giga 2.2 pro and ptgui 8.3.10 were used to stitch or mosaic a number of images captured by the cropcam uav platform. to cover a larger area, each flight's raw images were mosaicked together. to achieve a good mosaic, a couple of control points shared by photos taken in succession were manually added. because it was pre-programmed into the fly files, the overlap between adjunct photos was ideal. this has enabled the 47 | international journal of informatics information system and computer engineering 2(2) (2021) 44-52 stitching of images into seamless final mosaics, as well as the improvement of the georeferencing process (mengxiao, et al., 2018). figure 2 depicts the mosaicking procedure used in this study. the seemliness between individual mosaic pieces can be placed manually or automatically to be as inconspicuous as possible, and radiometric matching techniques can be used to account for color and brightness differences (tian, et al., 2020). all of the image mosaics in this study were created with the software packages autopano and ptgui. the original images (raw images) from each flight were mosaicked with lens distortion correction and color equalization. as shown in figure 2, a near-neighbor interpolator and smart blend bending algorithms were used to render images in order to make the mosaics shown in figure 2. fig. 2. modular workflow for image mosaicking process. 3. results and discussion figures 3 and 4 show examples of unprocessed digital images collected during cropcam uav flight missions over the chosen study sites. figures 5 and 6 depict image mosaics created from raw images collected after each flight over the study sites. the rmse of image mosaicking is displayed in table 1. faez m. hassan, hussein abdelwahab mossa. image mosaicking using low-distance...| 48 fig. 3. samples of cropcam raw images (first flight on june 20th, 2011) fig. 4. samples of cropcam raw images (second flight on december 12th, 2011) 202011). 49 | international journal of informatics information system and computer engineering 2(2) (2021) 44-52 fig. 5. uncontrolled image mosaic of the selected area in penang island created with autopano pro and ptgui software from 65 images taken by cropcam uav on june 20th, 2011 (first flight) fig. 6. uncontrolled image mosaic of the selected area in penang island created with autopano pro and ptgui software from 86 images taken by cropcam uav on december 12th, 2011 (second flight). faez m. hassan, hussein abdelwahab mossa. image mosaicking using low-distance...| 50 table 1. mosaicking process results flight mission number of stitched images panorama fov rmse (cm) quality status panorama 100 dpi(m) first flight 65 68.03º×51.31º 2.6 v. good 3.57×2.69 second flight 86 79.71º×40.10º 2.1 v. good 3.76×1.59 an examination of the generated image mosaics reveals that they are preferable. clearly, the characteristics (roads, buildings, etc.) in those images are joined perfectly with the minimum distortion. furthermore, due to the efficient method (image blending) used to create a mosaic of high quality, those image mosaics are in good enough shape to be used in further image analysis. in uav-acquired images, radiometric variations of overlapping views are common. as a result, each image region retains its own color, brightness, and contrast during the image blending process. therefore, these overlapping regions blend into one another with no discernible pattern. there is sufficient evidence in this study to show that the cropcam flight missions were successful to obtain the desired images with high resolution, and the mosaicking results are visually pleasing. image mosaics frequently reveal differences in exposure across or between photographs. finally, color matching between the stitched images is required to hide the seams. clearly, the mosaicking results show that the software used in this study can be used to solve the problem of uneven brightness in mosaics. 4. conclusion this research paper describes a simple but effective procedure for uav high-resolution mosaicking images obtained from images captured by uav flying at a low distance. the proposed method's performance was demonstrated in a case study on penang island, malaysia. the method proposed outperformed the highly developed commercial software on uav images to achieve better mosaicking results. our proposed method generates mosaicked images with reduced spectral distortion and increased spatial accuracy. moreover, the process of mosaicking is much faster than that of other software packages. the rmse value in the experimental results is quite high. the method presented in this paper saves 40% of the time. furthermore, the mosaicked images generated by our proposed technique are strikingly similar to the original uav images. in this paper, we propose a small-scale uav-based system for creating low-altitude image mosaics that are incremental and georeferenced in real time. acknowledgments we acknowledge mustansiriyah university, baghdad, iraq and universitas komputer indonesia, indonesia. 51 | international journal of informatics information system and computer engineering 2(2) (2021) 44-52 references aber, j. s., marzolff, i., & ries, j. (2010). small-format aerial photography: principles, techniques and geoscience applications. elsevier. ahmad, a. (2011). digital mapping using low altitude uav. pertanika journal of science and technology, 19(s), 51-58. avola, d., cinque, l., foresti, g. l., martinel, n., pannone, d., & piciarelli, c. (2018). a uav video dataset for mosaicking and change detection from low-altitude flights. ieee transactions on systems, man, and cybernetics: systems, 50(6), 2139-2149.. cropcam, (2008). cropcam user’s guide-application version, canada. felderhof, l., gillieson, d., zadro, p., & van boven, a. (2008). linking uav (unmanned aerial vehicle) technology with precision agriculture. fuyi, t., chun, b. b., jafri, m. z. m., san, l. h., abdullah, k., & tahrin, n. m. (2012, november). land cover/use mapping using multi-band imageries captured by cropcam unmanned aerial vehicle autopilot (uav) over penang island, malaysia. in unmanned/unattended sensors and sensor networks ix (vol. 8540, pp. 147-152). spie. gomarascs, m. a. (2009). basics of geomatics. london and new york, springer. gómez-candón, d., de castro, a. i., & lópez-granados, f. (2014). assessing the accuracy of mosaics from unmanned aerial vehicle (uav) imagery for precision agriculture purposes in wheat. precision agriculture, 15(1), 44-56. hassan, f. m., lim, h. s., & jafri, m. m. (2011). cropcam uav for land use/land cover mapping over penang island, malaysia. pertanika journal of science & technology, 19(s), 69-76. hassan, f. m., lim, h. s., matjafri, m. z., & othman, n. (2010, july). an assessment of low‐cost cropcam uav images for land cover/use over penang island, malaysia. in aip conference proceedings (vol. 1250, no. 1, pp. 23-26). american institute of physics. kim, j. i., kim, t., shin, d., & kim, s. (2017). fast and robust geometric correction for mosaicking uav images with narrow overlaps. international journal of remote sensing, 38(8-10), 2557-2576. lim, h. s., jafri, m. z. m., abdullah, k., hassan, f., & saleh, n. m. (2009). feasibility of using multi-band imageries captured by cropcam unmanned aerial vehicle autopilot for land cover mapping. journal of materials science and engineering, 3(12), 26-31. faez m. hassan, hussein abdelwahab mossa. image mosaicking using low-distance...| 52 mengxiao song, zheng ji, shan huang & jing fu. (2018). mosaicking uav orthoimages using bounded voronoi diagrams and watersheds. international journal of remote sensing, 39(15-16), 4960-4979. neteler, m., & mitasova, h. (2013). open source gis: a grass gis approach (vol. 689). springer science & business media. ren, x., sun, m., zhang, x., & liu, l. (2017). a simplified method for uav multispectral images mosaicking. remote sensing, 9(9), 962. tian, y., sun, a., luo, n., & gao, y. (2020). aerial image mosaicking based on the 6dof imaging model. international journal of remote sensing, 41(1), 74-89. wang, p. and xu, y., (2007). photogrammetry, wuhan: wuhan university press, pp.16 17 . xu, y., ou, j., he, h., zhang, x., & mills, j. (2016). mosaicking of unmanned aerial vehicle imagery in the absence of camera poses. remote sensing, 8(3), 204. zhao, j., zhang, x., gao, c., qiu, x., tian, y., zhu, y., & cao, w. (2019). rapid mosaicking of unmanned aerial vehicle (uav) images for crop growth monitoring using the sift algorithm. remote sensing, 11(10), 1226. zhou, g., wu, j., wright, s., & gao, j. (2006). high-resolution uav video data processing for forest fire surveillance. old dominion univ., norfolk, va, tech. rep. national sci. foundation. 31 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 unique aspects of usage of the quadratic cryptanalysis method to the gost 28147-89 encryption algorithm bardosh akhmedov*, rakhmatillo aloev** university of uzbekistan named after mirzo ulugbek tashkent, uzbekistan *corresponding email: shirin07@ya.ru a b s t r a c t s a r t i c l e i n f o in this article, issues related to the application of the quadratic cryptanalysis method to the five rounds of the gost 28147-89 encryption algorithm are given. for example, the role of the bit gains in the application of the quadratic cryptanalysis method, which is formed in the operation of addition according to mod232 used in this algorithm is described. in this case, it is shown that the selection of the relevant bits of the incoming plaintext and cipher text to be equal to zero plays an important role in order to obtain an effective result in cryptanalysis. article history: received 18 dec 2022 revised 20 dec 2022 accepted 25 dec 2022 available online 26 dec 2022 aug 2018 __________________ keywords: gost 28147-89, selected plaintext, quadratic approximation, correlation matrix, quadratic cryptanalysis 1. introduction in order to verify and evaluate the strength of encryption algorithms the possibilities of linear, differential, lineardifferential, algebraic, and correlation cryptanalysis are used. many works are devoted to improving applications of linear cryptanalysis. several linear approximations simultaneously for one combination of key bits (kaliski & robshaw, 1994; quisquater, 2004) can be used to increase the efficiency of the linear cryptanalysis method. a method for improving the lc method (in particular, for the cipher loki91) is proposed, which suggests taking into account the probabilistic behavior of some bits instead of their fixed values when approximating (sakurai & furuya, 1997). 2. literature review 2.1. linear cryptanalysis a series of works is devoted to the issues of the resistance of various encryption algorithms to the linear cryptanalysis method. in (chee et al., 1994), l.knudsen considered the issues of constructing feistel-type encryption schemes that are resistant to linear and differential international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 31-40 bardosh akhmedov and rakhmatillo aloev. unique aspects of usage of the quadratic … | 32 cryptanalysis methods. v.shorin, v.zheleznyakov and e.gabidulin proved in 2001 that the russian algorithm gost 28147-89 is resistant to these methods (with no less than five rounds of encryption in linear cryptanalysis and seven rounds in a different one). a large number of works are devoted to the study of various classes of approximating functions and to the construction of functions that are most difficult to such approximations. in these papers, bent functions (logachev et al., 2004; dobbertin & leander, 2004; chee et al., 1994) are considered, which are boolean functions from an even number of variables that are maximally distant from the set of all linear functions in the hamming metric, as well as their generalizations: semi-bent functions (dobbertin & leander, 2005), partially bent functions (qu et al., 2000), z−bent functions (pfitzmann, 2003), homogeneous bent functions (kuzmin et al., 2006), hyper best functions (carlet & gaborit, 2006; youssef, 2007; kuz'min et al., 2008; knudsen & robshaw, 1996). the main idea of using linear cryptanalysis of nonlinear approximations (knudsen & robshaw, 1996) is to enrich the class of approximating functions (of m variables) with nonlinear functions and increase the quality of approximation due to this. in this case, the cryptanalyst has to deal with the difficulties of choosing nonlinear approximations and combining nonlinear approximations of individual rounds. 2.2. gost 28147-89 encryption algorithms in the gost 28147-89 encryption algorithm (kuryazov et al., 2017; vinokurov) mod232 addition operation is used, and this operation, by its nature, the value of each resulting bit is connected to the values of the incoming bits below it in order. the mathematical model of each bit of the result of this operation can be expressed as follows: 2mod)( 323232 kxp += ; 2mod)( 32313131 qkxp ++= ; 2mod)( 3222 qkxp ++= 2mod)( 2111 qkxp ++= the general mathematical model of addition operation according to mod232 can be expressed as follows (kuryazov et al., 2017): 0,1...32,2mod)( 331 ==++= + qiqkxp iiii (1) here, qi –addition of sum of all i-bits. in this case, when applying the linear cryptanalysis method, considering the influence of the bit in each position of the block to be reflected with the output bits, the problem of building a boolean function for each bit of the result of the addition operation according to mod232 was considered. an overview of this function is as follows (kuryazov et al., 2017). )2.(0,1...32 ),(, 33 11 == == ++ qi kxqkxqqkxp iiiiiiiiii based on the results of the research on the mod232 addition operation used in the gost 28147-89 encryption algorithm, the schematic view of one round of this 33 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 algorithm is as follows (fig. 1) (kuryazov et al., 2017). 3. method 3.1. a. quadratic relations of a special form in previous works, correlation matrix values for linear and quadratic dependences and appropriate approximation equations with probability r=7/8 were obtained for gost 28147-89 algorithm s box. these equations are effectively used in linear cryptanalysis to find key bits with high probability (akhmedov & aloev, 2020; akhmedov, 2021). these equations are shown in table 1. fig. 1. schematic view of one round of gost 28147-89 encryption algorithm 3.2. quadratic cryptanalysis with these approximation equations, a modification of the gost 28147-89 algorithm, that is, using the xor operation instead of the mod232 addition operation, was used for the 5th round of quadratic cryptanalysis, and the corresponding results were obtained (akhmedov & aloev, 2020; akhmedov, 2021). based on the quadratic cryptanalysis conducted for the 5th round of the gost 28147-89 algorithm, the addition operation according to mod232 is used for the s block reflections, when conducting cryptanalysis based on the correlation matrices of linear and quadratic connections, the fact that some bits of the plaintext and ciphertext are equal to zero ensures the formation of an effective approximation relationship. 4. results and discussion based on the concepts presented above, the quadratic dependence approximation equations in table 1 determined for the correlation matrices for the s3-block are analyzed. 1-round: for s3 block 〖(p〗 _1⨁p_3)(p_2⨁p_4)⨁p_1⨁p_3=c_3 with probability p=12/16 the position of variables in the round reflection of the approximation equality, according to 11bit left cyclic shift and addition of leftside appropriate bits {( p(41) ⊞ k1(9))⨁(p(43) ⊞ k1(11))}*{(p(42) ⊞ k1(10))⨁(p(44) ⊞ k1(12))} ⨁ (p(41) ⊞ k1(9))⨁(p(42) ⊞ k1(10)) =y1(32)⨁p(32) will have the form. in order not to encounter the problem of addition from the sum of bits in this equality, it is necessary to choose plaintexts that satisfy the condition p(42)=p(43)=p(44)=p(45)=0. since the addition of the sum of p(41) and k1(9) in this block does not affect equality, it can bardosh akhmedov and rakhmatillo aloev. unique aspects of usage of the quadratic … | 34 be obtained in the form p(41) ⊞k1(9)=p(41)⨁k1(9). in this case (p(41)⨁k1(9)⨁k1(11))*(k1(10)⨁k1(12)) ⨁ p(41)⨁k1(9)⨁k1(10))=y1(32)⨁ p(32) (3) the equation is formed. table 1. cloud users’ responsibility types of attacks № approximation equations p o ss ib il it y e x cl u si o n s1 𝑝1⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐3𝑐4⨁𝑐1⨁𝑐4 𝑝1𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝2𝑝4⨁𝑝2⨁𝑝4 = 𝑐1𝑐3⨁𝑐1𝑐4⨁𝑐2𝑐3⨁𝑐2𝑐4⨁𝑐1⨁𝑐3 𝑝1𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝2𝑝4⨁𝑝1⨁𝑝3 = 𝑐2⨁𝑐3⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝1⨁𝑝3 = 𝑐2⨁𝑐3⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝1⨁𝑝3 = 𝑐1𝑐3⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐2𝑐4⨁𝑐1⨁𝑐3⨁1 p = 1/8 ∆=3/4 s2 𝑝1⨁𝑝3⨁𝑝4 = 𝑐1⨁𝑐4⨁1 𝑝1⨁𝑝2 = 𝑐1⨁𝑐2⨁𝑐4 𝑝1⨁𝑝2⨁𝑝3 = 𝑐1𝑐3⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐2𝑐4⨁𝑐2⨁𝑐3 𝑝1𝑝2⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝3𝑝4⨁𝑝2⨁𝑝3 = 𝑐1⨁𝑐2⨁𝑐4 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝4⨁𝑝3 = 𝑐3⨁𝑐4⨁1 p = 1/8 ∆=3/4 35 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 s3 𝑝1⨁𝑝4 = 𝑐1⨁𝑐3 𝑝2⨁𝑝3⨁𝑝4 = 𝑐1⨁𝑐3⨁𝑐4 𝑝2⨁𝑝3 = 𝑐1⨁𝑐2⨁𝑐3⨁1 𝑝1⨁𝑝3 = 𝑐1𝑐2⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐3𝑐4⨁𝑐1⨁𝑐2⨁1 𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐2⨁𝑐4⨁1 𝑝1⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐3⨁𝑐4 𝑝1𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝2𝑝4⨁𝑝1⨁𝑝3 = 𝑐1⨁𝑐2⨁1 𝑝1𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝2𝑝4⨁𝑝2⨁𝑝4 = 𝑐1⨁𝑐3 𝑝1𝑝2⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝3𝑝4⨁𝑝1⨁𝑝4 = 𝑐1⨁𝑐2⨁𝑐3 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝1⨁𝑝3 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐1⨁𝑐3⨁1 𝑝1𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝2𝑝4⨁𝑝1⨁𝑝3 = 𝑐1𝑐2⨁𝑐1𝑐3⨁𝑐2𝑐4⨁𝑐3𝑐4⨁𝑐3⨁𝑐4 p = 1/8 ∆=3/4 s4 𝑝3 = 𝑐1⨁𝑐2⨁𝑐3⨁𝑐4⨁1 𝑝1⨁𝑝3⨁𝑝4 = 𝑐1 𝑝2⨁𝑝4 = 𝑐3⨁1 𝑝3⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐3𝑐4⨁𝑐2⨁𝑐3 𝑝3⨁𝑝4 = 𝑐1𝑐2⨁𝑐1𝑐3⨁𝑐2𝑐4⨁𝑐3𝑐4⨁𝑐1⨁𝑐3 𝑝1⨁𝑝2⨁𝑝3⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐1⨁𝑐2⨁1 𝑝1⨁𝑝2⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐3⨁𝑐4 𝑝1𝑝2⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝3𝑝4⨁𝑝2⨁𝑝3 = 𝑐4⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝2⨁𝑝4 = p = 1/8 ∆=3/4 bardosh akhmedov and rakhmatillo aloev. unique aspects of usage of the quadratic … | 36 𝑐3⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝1⨁𝑝2 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐1⨁𝑐3⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝1⨁𝑝2 = 𝑐1𝑐3⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐2𝑐4⨁𝑐1⨁𝑐3⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝3⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐2⨁𝑐4 s5 𝑝3 = 𝑐1⨁𝑐2⨁𝑐3⨁𝑐4 𝑝1⨁𝑝2⨁𝑝3⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐3𝑐4⨁𝑐1⨁𝑐2 𝑝1𝑝2⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝3𝑝4⨁𝑝1⨁𝑝4 = 𝑐1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝2⨁𝑝4 = 𝑐1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝3⨁𝑝4 = 𝑐1⨁𝑐2⨁𝑐3⨁1 p = 1/8 ∆=3/4 s6 𝑝3⨁𝑝4 = 𝑐1⨁𝑐3⨁𝑐4 𝑝1⨁𝑝2⨁𝑝3 = 𝑐3 𝑝3 = 𝑐1𝑐3⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐2𝑐4⨁𝑐2⨁𝑐3⨁1 𝑝1⨁𝑝2⨁𝑝4 = 𝑐1𝑐2⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐3𝑐4⨁𝑐2⨁𝑐3⨁1 𝑝3 = 𝑐1𝑐2⨁𝑐2𝑐4⨁𝑐1𝑐3⨁𝑐3𝑐4⨁𝑐1⨁𝑐2⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝2⨁𝑝4 = 𝑐1⨁𝑐3⨁𝑐4 p = 1/8 ∆=3/4 s7 𝑝2⨁𝑝3⨁𝑝4 = 𝑐2⨁𝑐4 𝑝1⨁𝑝4 = 𝑐1𝑐3⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐2𝑐4⨁𝑐1⨁𝑐3⨁1 p = 1/8 ∆=3/4 37 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 𝑝3⨁𝑝4 = 𝑐1𝑐2⨁𝑐1𝑐4⨁𝑐2𝑐3⨁𝑐3𝑐4⨁𝑐1⨁𝑐2⨁1 𝑝1𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝2𝑝4⨁𝑝1⨁𝑝3 = 𝑐2⨁𝑐3⨁1 𝑝1𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝3⨁𝑝2𝑝4⨁𝑝2⨁𝑝4 = 𝑐3 𝑝1𝑝2⨁𝑝2𝑝3⨁𝑝2𝑝3⨁𝑝3𝑝4⨁𝑝1⨁𝑝2 = 𝑐2⨁𝑐3⨁1 𝑝1𝑝2⨁𝑝1𝑝3⨁𝑝2𝑝4⨁𝑝3𝑝4⨁𝑝2⨁𝑝4 = 𝑐3 𝑝1𝑝3⨁𝑝2𝑝3⨁𝑝1𝑝4⨁𝑝2𝑝4⨁𝑝1⨁𝑝3 = 𝑐1𝑐3⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐2𝑐4⨁𝑐1⨁𝑐3⨁1 s8 𝑝1⨁𝑝3 = 𝑐1⨁𝑐4⨁1 𝑝2 = 𝑐1⨁𝑐2 𝑝2⨁𝑝3 = 𝑐2⨁𝑐3⨁𝑐4⨁1 𝑝1⨁𝑝2⨁𝑝4 = 𝑐1⨁𝑐2⨁𝑐3⨁𝑐4⨁1 𝑝1⨁𝑝3 = 𝑐1𝑐2⨁𝑐2𝑐3⨁𝑐1𝑐4⨁𝑐3𝑐4⨁𝑐2⨁𝑐3 𝑝2 = 𝑐1𝑐2⨁𝑐1𝑐4⨁𝑐2𝑐3⨁𝑐3𝑐4⨁𝑐1⨁𝑐2 𝑝1⨁𝑝4 = 𝑐1𝑐2⨁𝑐1𝑐4⨁𝑐2𝑐3⨁𝑐3𝑐4⨁𝑐3⨁𝑐4 p = 1/8 ∆=3/4 2-round: in block s8, the equality p_4=c_1⨁c_2⨁c_4⨁1 has probability p=12/16, the variables in the approximation equality have the appearance p2(32)⊞k2(32)=y2(18)⨁y2(19)⨁y2(21) ⨁ p(18)⨁p(19)⨁p(21)⨁1 according to the position of the round reflection, a cyclic left shift of 11 bits, and the addition of left-side appropriate bits. since the value p2(32) in this parity represents the last bit, it is not affected by the summation, and since no other incoming text and key bits are involved, the addition p2(32) ⊞ k2(32) is not involved. in this case, equation p2(32) ⨁ k2(32) =y2(18)⨁ y2(19)⨁ y2(21)⨁p(18) ⨁p(19) ⨁p(21) ⨁1 (4) will appear y1(32) = p2(32) as a result of combining equations 3 and 4 according to compatibility, the following equation results: bardosh akhmedov and rakhmatillo aloev. unique aspects of usage of the quadratic … | 38 y2(18)⨁y2(19)⨁y2(21)=(p(41)⨁k1(9)⨁ k1(11))*(k1(10)⨁k1(12))⨁p(41)⨁k1(9)⨁ k1(10))⨁p(32)⨁p(18)⨁p(19)⨁p(21)⨁ k2(32)⨁1 (5) 3-round: in s3 block 〖(p〗 _1⨁p_3)(p_2⨁p_4)⨁p_1⨁p_3=c_3 with probability p=12/16 the variables in the approximation equality have the appearance of {(с(41) ⊞k5(9))⨁(с(43) ⊞k5(11))}*{(с(42) ⊞k5(10))⨁(с(44) ⊞ k5(12))} ⨁(с(41) ⊞ k5(9))⨁(с(42) ⊞ k5(10)) =y5(32)⨁с(32) according to the position in the round reflection, 11-bit left cyclic shift and addition of the left appropriate bits. in order not to encounter the problem of addition from the sum of bits in this equality, it is necessary to choose plaintexts that satisfy the condition s(42)=s(43)=s(44)=s(45)=0. since the addition of the sum of s(41) and k5(9) in this block does not affect equality, it can be obtained in the form s(41) ⊞k5(9)=s(41)⨁k5(9). in this case (s(41)⨁k5(9)⨁k5(11))*(k5(10)⨁k5(12)) ⨁s(41)⨁ k5(9)⨁k5(10))=y5(32)⨁s(32) (6) results in equality. 4-round: in block s8, the equality p_4=c_1⨁c_2⨁c_4⨁1 has probability p=12/16, the variables in the approximation equality have the appearance of p4(32) ⊞ k4(32) =y4(18)⨁ y4(19)⨁y4(21) ⨁с(18) ⨁с(19) ⨁с(21) ⨁1 according to the position of the round reflection, 11-bit left cyclic shift and addition of the left-side appropriate bits. since the value p4(32) in this parity represents the last bit, it is not affected by the summation, and since no other incoming text and key bits are involved, the addition p4(32) ⊞ k4(32) is not involved. in this case p4(32) ⨁ k4(32) =y4(18)⨁y4(19)⨁y4(21)⨁s(18) ⨁s(19) ⨁s(21)⨁1 (7) equality is formed. according to y5(32)=p4(32), combining equations 6 and 7 results in the following equation: y4(18)⨁y4(19)⨁y4(21)=(c(41)⨁k5(9)⨁ k5(11))*(k5(10)⨁k5(12))⨁c(41)⨁k5(9) ⨁ k5(10))⨁c(32)⨁c(18)⨁c(19)⨁c(21)⨁ k4(32) ⨁1 (8) 5 and 8 equations y4(18)⨁y4(19)⨁y4(21)=y2(18)⨁y2(19) ⨁ y2(21) based on (p(41)⨁k1(9)⨁k1(11))*(k1(10)⨁k1(12)) ⨁p(41)⨁k1(9)⨁k1(10))⨁p(32)⨁p(18)⨁ p(19)⨁p(21)⨁2(32)⨁1=(c(41)⨁k5(9)⨁k 5(11))*(k5(10)⨁k5(12))⨁c(41)⨁k5(9)⨁ k5(10)) ⨁k4(32)⨁c(32)⨁c(18)⨁c(19)⨁c(21)⨁1 and this results the following: (p(41)⨁k1(9)⨁k1(11))*(k1(10)⨁k1(12)) ⨁k1(9)⨁k1(10)⨁k2(32)⨁(c(41)⨁k5(9) ⨁k5(11))* (k5(10)⨁k5(12))⨁c(41)⨁k5(9)⨁k5(10)) ⨁ k4(32)=p(32)⨁p(18)⨁p(19)⨁p(21)⨁c(32 )⨁ c(18)⨁c(19)⨁c(21) (9) 39 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 the problem with sum-of-bits does not arise due to the fact that parity in general satisfies the following conditions: { p(41) = p(42) = p(43) = p(44) = p(45) = 0 с(41) = с(42) = с(43) = с(44) = с(45) = 0 the solution to the above problem with sum-of-bits depends on the s-block being chosen and requires a different approach. 5. conclusion modification of the gost 28147-89 algorithm, that is, using the xor operation instead of the mod232 operation, results of quadratic cryptanalysis method for the 5th round was used based on addition of bits using mod232 addition operation. due to this operation, the number of unknowns in the equation increases, since the value of each resulting bit depends on the values of the bits preceding it in order. for this reason, it is desirable to choose zero values of the corresponding bits of data entering the first round and exiting the fifth round in order to achieve an efficient result. since the second and fourth round input bits depend on the output values from the first and fifth round reflections, there is no option to select these bits. for this reason, it is necessary to consider these values as unknown. the stages of using the quadratic cryptanalysis method for five rounds of gost 28147-89 algorithm are created. in order to achieve an effective result in this method, it is shown that it is important to select the zero values of the corresponding bits of data entering the first round and exiting the fifth round. references akhmedov b.b. “nonlinear cryptanalysis for modification of the xor encryption algorithm gost 28147-89”, i международная научно-практическая интернет-конференция «актуальные вопросы физикоматематических и технических наук: теоретические и прикладные исследования», г.киев. 2021 г. 81-97 стр. www.openscilab.org. akhmedov b.b., aloev r.d. application of quadratic cryptanalysis for a five round xor modification of the encryption algorithm gost 28147-89 // international journal of science and research (ijsr), https://www.ijsr.net/search_index_results_paperid.php?id=sr2081818033 5, volume 9 issue 8, august 2020, 1101 – 1109, issn: 2319-7064, india). carlet, c., & gaborit, p. (2006). hyper-bent functions and cyclic codes. journal of combinatorial theory, series a, 113(3), 466-482. chee, s., lee, s., & kim, k. (1994, november). semi-bent functions. in international conference on the theory and application of cryptology (pp. 105-118). springer, berlin, heidelberg. bardosh akhmedov and rakhmatillo aloev. unique aspects of usage of the quadratic … | 40 dobbertin, h., & leander, g. (2004, october). a survey of some recent results on bent functions. in international conference on sequences and their applications (pp. 1-29). springer, berlin, heidelberg. dobbertin, h., & leander, g. (2005). cryptographer's toolkit for construction of $8 $bit bent functions. cryptology eprint archive. kaliski, b. s., & robshaw, m. j. (1994, august). linear cryptanalysis using multiple approximations. in annual international cryptology conference (pp. 26-39). springer, berlin, heidelberg. knudsen, l. r., & robshaw, m. j. (1996, may). non-linear approximations in linear cryptanalysis. in international conference on the theory and applications of cryptographic techniques (pp. 224-236). springer, berlin, heidelberg. kuz’min, a. s., markov, v. t., nechaev, a. a., shishkin, v. a., & shishkov, a. b. (2008). bent and hyper-bent functions over a field of 2ℓ elements. problems of information transmission, 44(1), 12-33. kuzmin, a. s., markov, v. t., nechaev, a. a., & shishkov, a. b. (2006). approximation of boolean functions by monomial ones. logachev, o. a., sal’nikov, a. a., & yashchenko, v. v. (2004). boolean functions in coding theory and cryptology. mccme, moscow. pfitzmann, b. (ed.). (2003). advances in cryptology–eurocrypt 2001: international conference on the theory and application of cryptographic techniques innsbruck, austria, may 6–10, 2001, proceedings (vol. 2045). springer. qu, c., seberry, j., & pieprzyk, j. (2000). homogeneous bent functions. discrete applied mathematics, 102(1-2), 133-139. quisquater, b. a. d. c. c. (2004). m franklin m on multiple linear approximations. in advances in cryptology–crypto (vol. 2004). sakurai, k., & furuya, s. (1997, january). improving linear cryptanalysis of loki91 by probabilistic counting method. in international workshop on fast software encryption (pp. 114-133). springer, berlin, heidelberg. vinokurov, a. algorithm for cryptographic data transformation gost 28147 89. youssef, a. m. (2007). generalized hyper-bent functions over gf (p). discrete applied mathematics, 155(8), 1066-1070. кuryazov d.m., sattarov a.b., akhmedov b.b. блокли симметрик шифрлаш алгоритмлари бардошлилигини замонавий криптотаҳлил усуллари билан баҳолаш. ўқув қўлланма. т.: «aloqachi». 2017, 228 бет. 59 | international journal of informatics information system and computer engineering 3(1) (2022) 59-70 association analysis with apriori algorithm for electronic sales decision support system r. fenny syafariani mathematics department, faculty of ocean technology engineering and informatics, universiti malaysia terengganu, malaysia corresponding email: r.fenny.syafariani@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o the purpose of this study was to determine the level of dependence of various items in order to dig up information on what items are dependent on other items. the method used in this research is descriptive analysis with a qualitative approach through a priori algorithm. the results show that the association analysis of the 26 transactions taken is 76.47%. a consumer who buys a laptop electronic device has the possibility to also buy an electronic mouse. article history: received 25 may 2022 revised 30 may 2022 accepted 10 june 2022 available online 26 june 2022 aug 2018 __________________ keywords: technology, information system, apriori, algorithm, decision support system, electronic sales 1. introduction data mining is a data processing method to find patterns from the data obtained (ordila et al., 2020). there are many methods in data mining. one method that is often used is the association method or association rule, more precisely using the apriori algorithm. the data generated from the sales process or transaction data is processed by association rules to find out information related to product purchases made by buyers (riszky et al., 2019). there are various kinds of electronic goods that are sold such as laptops, printers, mouse and so on. sales transactions continue to grow every day and cause huge data storage (purnia et al., 2017). most sales transaction data is only used as an archive without being international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 59-70 mailto:r.fenny.syafariani@email.unikom.ac.id r. fenny association analysis with apriori...| 60 used properly. however, this data set contains very useful information. with the application of association analysis or association rule mining in this discussion, it is hoped that association rules can be found between a combination of items. so that obtained a knowledge of the application of the concept of association mining analysis through the search for support and confidence. the previous research discussed "application of data mining for analysis of consumer purchase patterns with the fpgrowth algorithm on motorcycle spare part sales transaction data" (fajrin et al., 2018). this research is tested in order to influence consumer buying patterns, because each consumer's buying pattern is different. this needs to be analyzed further so that it can produce useful information, as well as maximize the benefits that can be obtained. then the next research that has been done previously is discussing "data mining analysis for clustering covid-19 cases in lampung province with the k-means algorithm" (nabila et al., 2021). this study was to analyze data on covid-19 cases in order to find out the grouping of the covid-19 case problems in lampung province. the grouping of data on covid19 cases in lampung province was carried out using the clustering method with the k-means algorithm. the results of dbi validation using manual calculations and using the help of rapidminer tools have differences, in this case manual calculations have better results than using rapidminer tools, but the results of both calculations are both close to 0 which means the evaluated clusters produce good clusters. in the previous research conducted using the fpgrowth algorithm analysis method, and subsequent research using the k-means method in conducting the analysis. this research is "association analysis with apriori algorithm for decision support system for selling electronic goods" (riszky et al., 2019). data mining and a priori algorithms are very useful to find out the reality of the frequency of sales of electronic goods that are most in demand by consumers. so that it can be used as very useful information in making decisions to prepare stocks of what types of electronic goods are needed in the future. 2. method in this study using descriptive analysis method with a qualitative approach (rahmawati et al., 2018). while in data processing using data mining techniques. the algorithm approach used is the a priori algorithm. the process of forming a combination of itemsets pattern and making rules starts from data analysis. the data used is data on sales of electronic goods, then followed by the formation of a combination of itemsets pattern and from an interesting combination of itemsets, association rules are formed. then the data is made in tabular data format (tana et al., 2018; syahril et al., 2020; simbolon, 2019). in relation to the application used in the test, it is an application that uses one of the microsoft excel databases with data in tabular data, then the sales transaction data (electronic goods data out), is converted into binary form (triansyah et al., 2018). after that the formation of a combination of two elements with a minimum value of the frequency of occurrence = 15 and a minimum value of confidence = 75%. to 61 | international journal of informatics information system and computer engineering 3(1) (2022) 59-70 calculate support and confidence, the following formula is used: 3. results and discussion 3.1. preprocessing the dataset used as a test sample in this study uses 26 transaction data. in the data there are several items of electronic goods sold, namely printers, laptops, chargers, and mice. the following is a table of transaction data that is used as a sample. table 1. transaction data table transaction item e1 printer,laptop e2 laptop, printer e3 charger, printer e4 printer, mouse, printer e5 charger, printer,printer e6 laptop, printer e7 printer,laptop,charger e8 printer, laptop, printer e9 printer,laptop e10 printer, printer e11 printer, laptop, printer e12 laptop, printer e13 charger, printer,printer e14 printer,laptop e15 printer,laptop,charger e16 printer, charger e17 charger, printer e18 charger, printer,printer e19 printer, laptop, printer support = 𝛴𝑖𝑡𝑒𝑚𝑠 𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒𝑑 𝑎𝑡 𝑜𝑛𝑐𝑒 𝛴𝑡𝑜𝑡𝑎𝑙 𝑡𝑟𝑎𝑛𝑠𝑎𝑐𝑡𝑖𝑜𝑛 × 100% confidence = 𝛴𝑖𝑡𝑒𝑚𝑠 𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒𝑑 𝑎𝑡 𝑜𝑛𝑐𝑒 𝛴𝑡𝑟𝑎𝑛𝑠𝑎𝑐𝑡𝑖𝑜𝑛𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑎𝑛𝑡𝑒𝑐𝑒𝑑𝑒𝑛𝑡 𝑠𝑒𝑐𝑡𝑖𝑜𝑛 × 100 % r. fenny association analysis with apriori...| 62 e20 charger, printer,printer e21 laptop, printer e22 printer,laptop,charger e23 printer,laptop e24 printer, mouse, printer e25 laptop, printer e26 printer, mouse, printer 3.2. transaction data tabular format the application used in the test is a microsoft excel database so that the data must be converted into binary form (sianturi et al., 2018).the conversion process is that the slip number of the data to be tested is made horizontally downwards, while all types of items will become vertical attributes, so that they form like a table, based on real sales data (electronic goods data out) the meeting point between the name of the electronic type and the number slip will become binary 1, while those that do not become meeting points will become binary 0. the results of the conversion process of sales transaction data to data format in tabular data form are as shown in the following table: table 2. data in the form of tabular data transaction printer mouse laptop charger e1 1 0 1 0 e2 0 1 1 0 e3 1 0 0 1 e4 1 1 1 0 e5 1 1 0 1 e6 0 1 1 0 e7 0 1 1 1 e8 1 1 1 0 e9 1 0 1 0 e10 1 1 0 0 63 | international journal of informatics information system and computer engineering 3(1) (2022) 59-70 e11 1 1 1 0 e12 0 1 1 0 e13 1 1 0 1 e14 1 0 1 0 e15 0 1 1 1 e16 0 1 0 1 e17 1 0 0 1 e18 1 1 0 1 e19 1 1 0 0 e20 1 1 0 1 e21 0 1 1 0 e22 0 1 1 1 e23 1 0 1 0 e24 1 1 1 0 e25 0 1 1 0 e26 1 1 1 0 17 20 17 10 in the tabular data table above, the number of occurrences (electronic items that come out) of each item is: printer = 17, mouse = 20, laptop = 17, and charger = 10. 3.3. formation of two elements combination pattern with a minimum value of the frequency of occurrence ф= 15. in the form table tabular data, there is one electronic item that does not meet the provisions of the frequency limit, namely charger = 10, so in the formation of the pattern of these two elements we make a combination of pairs of 3 electronic items, namely printer-mouse, laptop printer, mouselaptop. the following are tables of 2 element combinations: table 3. two elements combination pattern (printer, mouse) transaction printer mouse f r. fenny association analysis with apriori...| 64 e1 1 0 s e2 0 1 s e3 1 0 s e4 1 1 p e5 1 1 p e6 0 1 s e7 0 1 s e8 1 1 p e9 1 0 s e10 1 1 p e11 1 1 p e12 0 1 s e13 1 1 p e14 1 0 s e15 0 1 s e16 0 1 s e17 1 0 s e18 1 1 p e19 1 1 p e20 1 1 p e21 0 1 s e22 0 1 s e23 1 0 s e24 1 1 p e25 0 1 s e26 1 1 p total (p) 11 table 4. two elements combination pattern (printer, laptop) 65 | international journal of informatics information system and computer engineering 3(1) (2022) 59-70 transaction printer laptop f e1 1 1 p e2 0 1 s e3 1 0 s e4 1 1 p e5 1 0 s e6 0 1 s e7 0 1 s e8 1 1 p e9 1 1 p e10 1 0 s e11 1 1 p e12 0 1 s e13 1 0 s e14 1 1 p e15 0 1 s e16 0 0 s e17 1 0 s e18 1 0 s e19 1 0 s e20 1 0 s e21 0 1 s e22 0 1 s e23 1 1 p e24 1 1 p e25 0 1 s e26 1 1 p total (p) 9 table 5. two elements combination pattern (printer, laptop) transaction mouse laptop f e1 0 1 s e2 1 1 p e3 0 0 s e4 1 1 p e5 1 0 s e6 1 1 p e7 1 1 p e8 1 1 p e9 0 1 s e10 1 0 s r. fenny association analysis with apriori...| 66 e11 1 1 p e12 1 1 p e13 1 0 s e14 0 1 s e15 1 1 p e16 1 0 s e17 0 0 s e18 1 0 s e19 1 0 s e20 1 0 s e21 1 1 p e22 1 1 p e23 0 1 s e24 1 1 p e25 1 1 p e26 1 1 p total (p) 13 from the tables of the 2 elements above, p means that the items are sold together, while s means that there are no items that are sold together or there is no transaction. σ represents the number of frequency items set. so that in the pattern of these two elements, the support value is obtained, namely: • printer – mouse = 11 • printer – laptop = 9 • mouse – laptop = 13 3.4. formation of three elements combination pattern the combination of the 2 elements in the table above, we can combine into 3 elements. for the set formed on these 3 elements are laptop, printer, mouse. here is a table of 3 elements: table 6. three elements combination pattern (printer, mouse, laptop) transaction printer mouse laptop f e1 1 0 1 s e2 0 1 1 s e3 1 0 0 s e4 1 1 1 p e5 1 1 0 s 67 | international journal of informatics information system and computer engineering 3(1) (2022) 59-70 e6 0 1 1 s e7 0 1 1 s e8 1 1 1 p e9 1 0 1 s e10 1 1 0 s e11 1 1 1 p e12 0 1 1 s e13 1 1 0 s e14 1 0 1 s e15 0 1 1 s e16 0 1 0 s e17 1 0 0 s e18 1 1 0 s e19 1 1 0 s e20 1 1 0 s e21 0 1 1 s e22 0 1 1 s e23 1 0 1 s e24 1 1 1 p e25 0 1 1 s e26 1 1 1 p total (p) 5 it can be seen from the pattern table of the 3 elements above, the items that were sold simultaneously were laptop – printer mouse = 5 so, the support value in the 3 element pattern table is 5. 3.4. association rules calculating the support and confidence values of each frequent itemset so that candidate association rules appear (lestari, 2017). to calculate support and confidence, the following formula is used: support = 𝛴𝑖𝑡𝑒𝑚𝑠 𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒𝑑 𝑎𝑡 𝑜𝑛𝑐𝑒 𝛴𝑡𝑜𝑡𝑎𝑙 𝑡𝑟𝑎𝑛𝑠𝑎𝑐𝑡𝑖𝑜𝑛 × 100% r. fenny association analysis with apriori...| 68 so that the results are obtained as in the table below table 7. association rules candidate list if antecedent, then consequent support confidence printer, mouse 11/26*100 %= 42,30% 11/17*100 %= 64,70% mouse, printer 11/26*100 %= 42,30% 11/20*100 %= 55% printer, laptop 9/26*100%=34,61% 9/17*100%=52,94% laptop, printer 9/26*100%=34,61% 9/17*100%=34,61% mouse, laptop 13/26*100%=50% 13/20*100%=65% laptop, mouse 13/26*100%=50% 13/17*100%=76,47% from the table above, the support and confidence have been determined. then select the association rules that meet the minimum confidence of 75%, so that the association rules are obtained, which are as follows: table 8. association rules list if antecedent, then consequent support confidence laptop, mouse 13/26*100%=50% 13/17*100%=76,47% from the results of the analysis that has been carried out, there is 1 product association rule that meets the minimum confidence limit, namely laptop mouse. then the results obtained are "76,47% of transactions that contain laptop electronics also contain mouse electronics. and 50% of all transactions that contain these two items". with apriori algorithm analysis can be applied to assist marketing strategies in a company or institutions. data mining and a priori algorithms are very useful to find out the relationship between the frequency of sales of electronic goods that are most in demand by customers, so that they can be used as very valuable information in making decisions to prepare stocks of what types of electronic goods are needed in the future. 4. conclusion a priori algorithm is used in conducting association analysis to determine the level of dependence of various items to explore information on what items have dependence on other items based on 26 transaction records that are sampled. the author performs an association analysis calculation from the samples taken so that the result is that 76.47% of a consumer who buys laptop electronics has the possibility to also buy mouse electronics. and 50% of all transactions that contain these two items. confidence = 𝛴𝑖𝑡𝑒𝑚𝑠 𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒𝑑 𝑎𝑡 𝑜𝑛𝑐𝑒 𝛴𝑡ℎ𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑟𝑎𝑛𝑠𝑎𝑐𝑡𝑖𝑜𝑛𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑎𝑛𝑡𝑒𝑐𝑒𝑑𝑒𝑛𝑡 𝑠𝑒𝑐𝑡𝑖𝑜𝑛 × 100 % 69 | international journal of informatics information system and computer engineering 3(1) (2022) 59-70 references ordila, r., wahyuni, r., irawan, y., & sari, m. y. (2020). penerapan data mining untuk pengelompokan data rekam medis pasien berdasarkan jenis penyakit dengan algoritma clustering (studi kasus: poli klinik pt. inecda). jurnal ilmu komputer, 9(2), 148-153. riszky, a. r., & sadikin, m. (2019). data mining menggunakan algoritma apriori untuk rekomendasi produk bagi pelanggan. jurnal teknologi dan sistem komputer, 7(3), 103-108. purnia, d. s., & warnilah, a. i. (2017). implementasi data mining pada penjualan kacamata menggunakan algoritma apriori. ijcit (indonesian journal on computer and information technology), 2(2). rahmawati, f., & merlina, n. (2018). metode data mining terhadap data penjualan sparepart mesin fotocopy menggunakan algoritma apriori. piksel: penelitian ilmu komputer sistem embedded and logic, 6(1), 9-20. triyansyah, d., & fitrianah, d. (2018). analisis data mining menggunakan algoritma k-means clustering untuk menentukan strategi marketing. incomtech: jurnal telekomunikasi dan komputer, 8(3), 163-182. nabila, z., isnain, a. r., permata, p., & abidin, z. (2021). analisis data mining untuk clustering kasus covid-19 di provinsi lampung dengan algoritma kmeans. jurnal teknologi dan sistem informasi, 2(2), 100-108. riszky, a. r., & sadikin, m. (2019). data mining menggunakan algoritma apriori untuk rekomendasi produk bagi pelanggan. jurnal teknologi dan sistem komputer, 7(3), 103-108. sianturi, f. a. (2018). penerapan algoritma apriori untuk penentuan tingkat pesanan. jurnal mantik penusa, 2(1). lestari, n. (2017). penerapan data mining algoritma apriori dalam sistem informasi penjualan. jurnal edik informatika penelitian bidang komputer sains dan pendidikan informatika, 3(2), 103-114. tana, m. p., marisa, f., & wijaya, i. d. (2018). penerapan metode data mining market basket analysis terhadap data penjualan produk pada toko oase menggunakan algoritma apriori. jimp-jurnal informatika merdeka pasuruan, 3(2). syahril, m., erwansyah, k., & yetri, m. (2020). penerapan data mining untuk menentukan pola penjualan peralatan sekolah pada brand wigglo dengan r. fenny association analysis with apriori...| 70 menggunakan algoritma apriori. jurnal teknologi sistem informasi dan sistem komputer tgd, 3(1), 118-136. simbolon, p. h. (2019). implementasi data mining pada sistem persediaan barang menggunakan algoritma apriori (studi kasus: srikandi cash credit elektronic dan furniture). jurikom (jurnal riset komputer), 6(4), 401-406. fajrin, a. a., & maulana, a. (2018). penerapan data mining untuk analisis pola pembelian konsumen dengan algoritma fp-growth pada data transaksi penjualan spare part motor. kumpulan jurnal ilmu komputer (klik), 5(01), 1-10. 203 | international journal of informatics information system and computer engineering 3(2) (2022) 209-218 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 log monitoring system using quick response (qr) code: a state university’s covid – 19 contact tracing system anna monica c. paculaba*, mark angelo s. bathan, erickson l. niego college of arts and sciences, samar state university, philippines *corresponding email: monica.paculaba@ssu.edu.ph a b s t r a c t s a r t i c l e i n f o contact tracing is the technique employed by public health units and the national close contact service to track down persons who may have been exposed to covid-19 by interaction with a suspect, confirmed, or probable case during their infectious period. this study focused on the development of a log monitoring system using quick response (qr) code in samar state university as an institution’s tracing system for covid – 19 preventions. the study was designed as a tool for managing the everyday logs of the employees, students, and visitors to track down the person who is in close contact to a covid – 19 positives. the waterfall model was used in developing the system and descriptive research design was used to determine the effectiveness of the system along with functionality, reliability, usability, efficiency, maintainability, and portability. the participants of the study were the employees, students, and visitors of ssu. each participant has given an iso 9126 quality standard questionnaire for the evaluation of the effectiveness of the system. the result revealed that using the system, the conduct of contact tracing of the possible covid – 19 suspected individuals was done easily and with reliability. article history: submitted/received 01 jul 2022 first revised 03 sept 2022 accepted 01 oct 2022 available online 13 oct 2022 publication date 01 dec 2022 aug 2018 __________________ keywords: information system, iso/iec 9126, monitoring system, system effectiveness. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 209-218 https://doi.org/10.34010/injiiscom.v3i2.9037 paculaba et al. log monitoring system using quick response (qr) …| 210 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 1. introduction during the current covid-19 pandemic and previous pandemics, a variety of digital health initiatives were used to control disease transmission. these control techniques have been shown to be effective in decreasing the initial wave of covid-19 in various countries; among these strategies, contact tracing is regarded the cornerstone of containment and receives a lot of attention. according to the study of ferretti, et al. (2020), contact tracing entails locating, quarantining, and notifying infected individuals' contacts. this is a technique employed by public health units and the national close contact service to track down persons who may have been exposed to covid 19 by interaction with a suspect, confirmed, or probable case during their infectious period. (new zealand government ministry of health, 2020). one of the most efficient approaches to stop the transmission of this virus and identify the main and secondary contacts of confirmed covid19 patients is to use this kind of technologies and tools (dunford, et al., 2020; chen, yang & wang, 2020). hence, it has led to a fast rise of covid-19 contact tracing technologies in different countries in the world. in the philippines, various contact tracing system has emerged. different local government units (lgus) in the different regions of the country implemented its contact tracing system to boost the lgus contact tracing program amid the rising number of covid – 19 cases. local government unit of catbalogan city, samar employed this type of system. it was deployed to the different establishments in the city to monitor the logs of the individuals using quick response (qr) code to capture easily their data. in this way, it will be accessible to locate the persons who are in close contact to a covid – 19 positive individuals. every resident of the city and non – residents who works in the said area need to register their information manually to avail a “quarantine pass” an identification card with qr code, that served as their access to enter different establishments. one of the establishments that was recipient of the said system is samar state university (ssu), one of the higher institutions in the philippines located in the city of catbalogan. the system was implemented in ssu, however, only those logs of the employees and visitors who have their quarantine passes can be recorded by the system. information of the students who came from other places outside catbalogan cannot be recorded. on the other hand, the system is limited only in capturing the logs who enter inside the campus. information of the individuals who visited the different departments, colleges, and offices in ssu were not included in the functionality of the system. as we know in this modern world, monitoring system is widely used in organizations like schools, to keep track of the day-today operations (balmes, 2016). on the other hand, using unique qr code-based identity card, authorization as well as authentication in monitoring is very important for the growth of the organization (bole, et al., 2016). hence, it is possessing a really great advantage that among the whole types of code scanning technology, qr code-based monitoring system is the most accurate (wei, et al., 2017). this led https://doi.org/10.34010/injiiscom.v3i2.9037 211 | international journal of informatics information system and computer engineering 3(2) (2022) 209-218 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 to an idea to the researchers to develop a web – based log monitoring system with an integration of qr code for samar state university. the system intends to quickly take a piece of information from the employees, students, and visitors who visited the campus in order to trace their daily logs that will be an advantage in tracking down the information of the persons who were in a close contact to a covid – 19 positives. individuals which include students, faculty and employees from ssu, and visitors from any part of the country who wanted to visit ssu can easily have their “quarantine pass” as the system has an automated registration. the system has the capability to record the entrance and exit logs inside the campus, and apart from that, information of an individual who enter and exit in the different department, colleges, and offices of ssu will be logged by the system. in this way, the duration of time of stay consumed by an individual in the different offices they visited within the campus can be easily tracked which was a big aspect to determine if a person is considered to be a close contact to a covid – 19 positive individuals. in addition, to test the effectiveness of the system, iso/iec 9126 system quality standard will be applied (iso/iec 9126, 2022). 2. method stakeholders in the ilorin metropolitan this study utilized a developmental approach in developing the log monitoring system using quick response (qr) code and used a quantitative approach to assess the system’s effectiveness upon implementation. the researchers have chosen the water fall system development life cycle model as shown in fig.1 in order to achieve the objectives of the study. the waterfall model defines several consecutive phases that must be completed one after the other and moving to the next phase only when its preceding phase is completely done. for this reason, the waterfall model is the waterfall model is recursive in that each phase can be endlessly repeated until it is perfected (bassil, 2012). this methodology is composed of five phases such as planning and analysis phase, design phase, implementation phase, testing phase and maintenance phase. 2.1. system planning and analysis the researchers determined the requirements such as gathering data from the end – user. after gathering of data, it was analyzed to test the validity. the probability of combining the requirements in the system to be developed is also studied. 2.2. system design the system was rationally planned to fulfill the functional requirements identified during analyzing requirements specification. in addition, the database as well as the interface was drawn on how it will be designed. 2.3. implementation in this phase, the whole system requirements and blue prints was converted into an actual coding. the researchers used the php programming language as a base program of the system and mysql for its database. 2.4. testing and integration testing of the product was made. each functionality was verified and validated by the researchers to evaluate if the vital specifications was met. https://doi.org/10.34010/injiiscom.v3i2.9037 paculaba et al. log monitoring system using quick response (qr) …| 212 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 2.5. operation and maintenance after testing and evaluating the system, the errors discovered from the deployment of the system was refined, corrected and modified as well as the suggestions gathered from the end – users was considered. 2.6. system architecture and flow the architecture and flow of the log monitoring system is presented in figs. 1 and 2. 2.7. data gathering procedure in general, the researchers provided a set of questionnaires as one of the basic instruments in conducting a beta-testing. with a total of forty (40) number of respondents; 10 faculty, 10 admin employees, 10 students, and 10 visitors to test the effectiveness of the system. the respondents tested the system’s effectiveness in terms of its accuracy, functionality, reliability, and efficiency. the researchers had also used applicable statistical tools for the complete evaluation of the system in order to acquire concrete feedbacks and opinions coming from of the respondents. fig. 1. developmental process fig. 2. system architecture and flow https://doi.org/10.34010/injiiscom.v3i2.9037 213 | international journal of informatics information system and computer engineering 3(2) (2022) 209-218 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 the effectiveness of the system applies the three (3) layers of testing namely: alpha, beta and full implementation. the mean uses the standard criteria under iso/iec 9126 in terms of functionality, reliability, usability, efficiency, maintainability and portability. 3. results and discussion anchored to the objectives of the study, the result of the study was derived. 3.1. system’s interface 3.1.1. log – in and registration before entering to the main setup form in the log monitoring system, a login form will pop up as shown in fig. 3 to allow the user to login with their own username and password. the function of the login form is to enter authentication credentials so that the users can access to the main form of the system. when the login form is submitted, the elemental code will be used to check and compare with the existing credentials in mysql database. if the result matched, the users will be granted for further features of the system. on the contrary, in order for the client to have an account in the log monitoring system, the client should register first by providing their personal information such as: first name, last name, contact number, address, client – type (faculty, student, admin employee, or visitor), and their respective offices/department. for the outsiders or the visitors, they need to select the offices/department that they wanted to visit. 3.1.2. homepage once the user successfully logged – in into the system, the user will be redirected to the homepage interface as shown in fig. 4. from the homepage, menu option is seen where qr code information is located. unique qr code is automatically generated, once the user provides their information and successfully access the system. 3.1.3. qr code scanner this interface shows how the qr code will be scanned as seen in fig. 5. once unique qr code is scanned, information of the user who enter the university will be automatically stored in the database. 3.1.4. log reports this log reports interface is shown in figs. 6 and 7. it presents the summary of the logs of the persons who enter and exit the campus. once the summary of log reports is downloaded, information of the person who enter the specific office, the time they log – in and log – out, contact number, address and their client type (faculty, student, admin employee, and visitor) displayed. reports can be downloaded in excel form. https://doi.org/10.34010/injiiscom.v3i2.9037 paculaba et al. log monitoring system using quick response (qr) …| 214 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 fig. 3. log – in and registration for fig. 4. homepage https://doi.org/10.34010/injiiscom.v3i2.9037 215 | international journal of informatics information system and computer engineering 3(2) (2022) 209-218 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 fig. 5. qr code scanner fig. 6. log report interface https://doi.org/10.34010/injiiscom.v3i2.9037 paculaba et al. log monitoring system using quick response (qr) …| 216 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 fig. 7. summary of log report 3.2. end user’s system performance evaluation the result of the end – user evaluation as shown in table 1 indicates that the system over – all performance is “highly effective” with a numerical mean of 4.61. among of the parameters, reliability of the system got the highest mean 4.68. the result implied that the system is free from error, capable of handling errors, can resume and restore lost data from failure, and can presents integrated reports. the result agreed to the study of tworek (2018), that the reliability of the information system is linked to the information security, availability, and responsiveness. this can be characterized as that the system is assured to be accurate and is conveniently accessible, user-friendly and accepted by its users, responsive, and high availability connected with high security. table 1. end user’s evaluation of system performance evaluation result criteria mean value descriptive remarks functionality 4.65 highly effective reliability 4.68 highly effective usability 4.58 highly effective efficiency 4.56 highly effective https://doi.org/10.34010/injiiscom.v3i2.9037 217 | international journal of informatics information system and computer engineering 3(2) (2022) 209-218 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 table 1 (continue). end user’s evaluation of system performance evaluation result maintainability 4.61 highly effective portability 4.55 highly effective grand mean 4.61 highly effective legend: 1.51 – 5:00 highly effective 3.31 – 4.50 moderately effective 2.51 – 3.50 effective 1.51 – 2.50 slightly effective 1:00 – 1.50 not effective 4. conclusion log monitoring system using qr code is one of the effective ways to monitor the logs of the faculty, students, admin employee, and visitors who enter the campus. since log – in and log – out of the respondents are recorded, the conduct of contact tracing of the possible covid – 19 suspected individuals will be done easily and with reliability as the span of time of their stay in the area/office inside the university they visited is recorded by the system. for the expansion and sustainable use of the log monitoring system using qr code, it is recommended that attendance monitoring of the students, laboratory utilization monitoring, and other related activities will be included in the system. acknowledgement the researchers would like to acknowledge the center for engineering, science and technology, and innovation (cesti) of samar state university for the supported funding they provided to make this research project possible references balmes, i. l. (2016). online day to day monitoring system for lyceum of the philippines university-batangas. the 2nd multidisciplinary research and innovation for globally sustainable development. bassil, y. (2012). a simulation model for the waterfall software development life cycle. international journal of engineering & technology (ijet), 2(5). bole, g. r., more, s. p., parnak, a. a., & naik, l. s. (2016). qr code based effective employee maintenance system. international research journal of engineering and technology (irjet), 03(04), 604 607. chen, s., yang, j., yang, w., wang, c., & bärnighausen, t. (2020). covid-19 control in china during mass population movements at new year. the lancet, 395(10226), 764-766. https://doi.org/10.34010/injiiscom.v3i2.9037 paculaba et al. log monitoring system using quick response (qr) …| 218 doi: https://doi.org/10.34010/injiiscom.v3i2.9037 p-issn 2810-0670 e-issn 2775-5584 dunford, d., dale, b., stylianou, n., lowther, e., ahmed, m., & de la torre arenas, i. (2020). coronavirus: the world in lockdown in maps and charts. bbc news, 9, 462. ferretti, l., wymant, c., kendall, m., zhao, l., nurtay, a., abeler-dörner, l., ... & fraser, c. (2020). quantifying sars-cov-2 transmission suggests epidemic control with digital contact tracing. science, 368(6491), eabb6936. iso/iec 9126 in software engineering. (2022). retrieved from greeks for greeks: https://www.geeksforgeeks.org/iso-iec-9126-in-software-engineering/ new zealand government ministry of health. (2020). retrieved from https://www.health.govt.nz/system/files/documents/publications/covi d-19_contact_tracing_qr_code_specification_25_may_2020.pdf tworek, k. (2018). reliability of information systems in organization in the context of banking sector: empirical study from poland. cogent business & management, 5(1), 1522752. wei, x., manori, a., devnath, n., pasi, n., & kumar, v. (2017). qr code based smart attendance system. international journal of smart business and technology, 5(1), 1-10. https://doi.org/10.34010/injiiscom.v3i2.9037 https://www.geeksforgeeks.org/iso-iec-9126-in-software-engineering/ https://www.health.govt.nz/system/files/documents/publications/covid-19_contact_tracing_qr_code_specification_25_may_2020.pdf https://www.health.govt.nz/system/files/documents/publications/covid-19_contact_tracing_qr_code_specification_25_may_2020.pdf 1 | international journal of informatics information system and computer engineering 4(1) (2023) 23-30 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 combination of technology acceptance model and decision-making process to study retentive consumer behavior on online shopping edwin rudini1*, dwiza riana2, sri hadianti3 1,2program studi ilmu komputer, universitas nusa mandiri 3program studi informatika, universitas nusa mandiri jln. jatiwaringin no.2 cipinang melayu, makasar jakarta timur, kota jakarta, 13620 dki jakarta *corresponding email: edwinrudini98@gmail.com a b s t r a c t s a r t i c l e i n f o during the spread of the covid-19 virus, generally the indonesian people began to switch from conventional markets to buying and selling goods and services online with various features and conveniences offered to users. the purpose of this study is to find out the extent to which indicators of satisfaction and trust influence consumer attitudes and behavior when deciding to make transactions at online shops. the study method uses a combination of tam (theory acceptance model) and dmp (decision making process) models using a sampling of 110 student respondents and the public who have made transactions in online shops. data analysis using sem (structural equation modeling) theory. the results showed that satisfaction and trust will influence consumers in shaping. article history: submitted/received 03 dec 2022 first revised 31 dec 2022 accepted 20 feb 2023 first available online 10 april 2023 publication date 01 june 2023 8 __________________ keywords: tam, dmp model, trust, satisfaction, repurchase. 1. introduction changing people's behavior within the framework of online stores is a big challenge for companies to be able to serve all people's needs and wants. information released by the ministry of communication and information explained that the value of online shopping transactions in 2021 will reach idr 337 trillion and the number of internet users will reach 210 million (rose et al., 2015). therefore, it can be concluded that the possibility of development in online trading is very open. this encourages several large companies to invest in the advancement of online business in indonesia. the growing potential of online commerce is international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(1) (2023) 23-30 https://doi.org/10.34010/injiiscom.v4i1.9414 mailto:edwinrudini98@gmail.com rudini et al., combination of technology acceptance model and decision-making … | 24 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 expected to produce more technology entrepreneurs and encourage the growth of msmes according to the characteristics of each company to utilize every potential they have (haryanti & subriadi, 2020). icd research foundation estimates the growth potential of online shop business in indonesia at 33.2% during 2020-2021 and is one of the countries with the fastest and largest online shopping-based business growth rate in the asia-pacific region (setyowati et al, 2021). the increase in online commerce activity is not in accordance with the growth of online shop buyers. this is due to a number of barriers, including low credit and debit card accessibility as well as shoppers' reluctance to shop online (nagy & hajdú, 2021). then based on nielsen statistical surveys it is known that buyers will look for data on the internet before choosing to buy a product they need. the trust to buy products in online shopping is a barrier difficult to control. because it is related to customer views and behavior (dennis et al., 2010). therefore, to study the attitude and behavior of buyers towards online shopping so that business organizations can take advantage of existing opportunities. it is possible to measure buyer behavior with a social behavior approach that acts as a variable that influences customer views and behavior in online shopping. to measure buyer behavior must be possible with a social behavior approach that acts as a variable that influences customer perspective and behavior in shopping at online shops (petcharat & leelasantitham, 2021). in estimating the use of data innovation, there are several models that can be used, such as the technology acceptance model (tam) and the decision making process (dmp), which states that individual behavior is a measure of power and activity whereby individuals will leverage data frameworks and data innovations if it is beneficial (wei et al., 2018). customer behavior in online commerce is also influenced by the satisfaction of making transactions on the internet and is a major factor that makes buyers prefer online stores. furthermore, buyer satisfaction with online transactions is proven to affect customer confidence, which in turn will affect buyers' views on repurchases. given the problems, obstacles and difficulties, as well as the potential created by online merchants, the online shopping sector is expected to encourage the improvement of the indonesian economy (kim, 2020). 2. method the study method uses a combination of tam (theory acceptance model) and dmp (decision making process) models. due to the large population and limited time and cost to 110 respondents from the population studied. in addition, research methods can be used to evaluate and compare results and draw conclusions. specimen collection techniques through targeted sampling of students and communities. in determining research specimens based on several criteria, namely students who are active, willing to answer surveys distributed by researchers, a minimum sample size of 15% of the total population, and have already done a set online shopping transaction (see tables 1 and 2). the method of collecting data is carried out by distributing questionnaires through google forms to respondents with specified randomization criteria and observing directly the object to be studied. in this study the variables used are independent variables, moderator https://doi.org/10.34010/injiiscom.v4i1.9414 25 | international journal of informatics information system and computer engineering 4(1) (2023) 23-30 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 variables and dependent variables. the independent variables used are perceived usefulness (x1), perceived ease of use (x2), evaluation alternative (x3), information search (x4), moderator variables namely trust (y1), satisfaction (y2) and the dependent variable used is purchase (z). table 1. respond demographics total (n-384) frequency percentage (%) gender male 24 26.8 female 86 73.20 nationality indonesia 110 100 18-35 81 29.17 age 36-53 15 18.08 54-65 14 17.05 table 2. variable and indicators variable indicators perceived usefullness (x1) online shopping platforms help you search and buy products faster than offline shopping online shopping platforms help you buy products cheaper than offline shopping perceived easy use (x2) you can use online shopping platforms with your own gadgets online shopping platforms have clear functions and are easy to understand evaluation alternative (x3) the search function in online shopping platforms is necessary and benefits shoppers intend to buy products before going to the shopping cart functions in online shopping platforms help you to compare a product information search (x3) the search function on online shopping platforms is quite helpful comparing the quality and price of similar products will make it easier for you to make a decision to buy perceived trust (y1) online shopping platforms have accurate and clear results such as product details and prices you are sure to get the purchased products from the online shopping platform satisfaction (y2) online shopping platforms have accurate and clear results such as product details and prices you are sure to get the purchased products from the online shopping platform an online shopping platform is required for you repurchase (z) you can buy back from online shopping platforms you have repurchased the same product from an online shopping platform https://doi.org/10.34010/injiiscom.v4i1.9414 rudini et al., combination of technology acceptance model and decision-making … | 26 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 3. results and discussion this analysis was used to describe the results of a survey consisting of the number of students and the general public who answered a questionnaire that measured their trust and satisfaction in using online shopping (bhatti et al., 2020). the data was processed using smart partial least squares (pls) software version 3.2.9 and microsoft excel windows 2016 version. the variables analyzed in this study include satisfaction before and after transactions (x1, x2, x3 and x4), trust (y1), and satisfaction of online shopping repurchases (y2) and purchases (z) (riantini, n.d.). the research instrumentation uses likert scale methodology. the likert scale consists of two types of statements, positive and negative, with positive statements scoring 4 points for strongly agreeable responses and strongly disagree responses as 1 point using structural equation modeling (sem) and pls data analysis techniques to develop predictive theories related to satisfaction and trust in the use of online shopping in students and the community (fedorko et al., 2018). pls model analysis is based on predictive measures with non-parametric properties due to convergence validity. here the measurement of individual reflections correlates with discriminant validity values comparing loading values of > 0.5 and squared values. root of extracted mean variance (ave) for each component with correlation between components in the model (sheth, 2020). discriminant validity is good if the ave value is greater than the correlation value between the component and the model. structural models are tested using r squared for dependent structures, stone geyser q-squared test to test predictive associations, t tests and significance for structural path parameters. data analysis was carried out by entering all respondents' data and testing convergence validity, discriminant validity and significance (cai et al., 2023). the results of the calculation explain that all indicators meet a construct loading value of >0.5 so that all indicators can be used in tests using the pls model. referring to the results of the calculation of convergence validity determined loading values per indicator is shown in table 3. table 3. convergent validity value evaluation alternative uses facilities belief satisfaction information search purchase ea1 0,879 ea2 0,830 is2 0,737 is3 0,941 peu1 1,000 ps1 0,899 ps2 0,897 pt1 0,920 pt2 0,880 pu2 1,000 rp2 1,000 https://doi.org/10.34010/injiiscom.v4i1.9414 27 | international journal of informatics information system and computer engineering 4(1) (2023) 23-30 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 table 4. discriminant validity evaluation alternative uses facilities belief satisfaction information search purchase evaluation alternativ e 0,855 uses 0,283 1,000 facilities 0,314 0,496 1,000 belief 0,344 0,377 0,152 0,900 satisfaction 0,497 0,455 0,510 0,353 0,898 informatio n search -0,099 0,167 0,272 -0,191 0,145 0,845 purchase 0,041 0,065 0,192 -0,069 -0,109 0,177 1,000 table 5. average variance extracted (ave) and composite reability cronbach's alpha rho_a composite reliability average variance extracted (ave) evaluation of alternativ e 0,633 0,644 0,844 0,731 uses 1,000 1,000 1,000 1,000 facilities 1,000 1,000 1,000 1,000 belief 0,769 0,789 0,896 0,811 satisfaction 0,760 0,760 0,893 0,806 informatio n search 0,634 0,845 0,831 0,714 purchase 1,000 1,000 1,000 1,000 table 6. path coefficient and decision original sample (o) sample mean (m) standard deviation (stdev) t statistics (|o/stdev|) p value evaluation alternative -> belief 0,232 0,257 0,099 2,331 0,020 evaluation alternativ e -> satisfaction 0,320 0,314 0,113 2,836 0,005 evaluation of alternativ e -> purchase 0,111 0,131 0,144 0,775 0,439 uses -> belief 0,371 0,352 0,104 3,554 0,000 uses -> satisfaction 0,143 0,146 0,094 1,511 0,132 uses -> purchase 0,051 0,055 0,158 0,325 0,745 facilities -> belief -0,046 -0,054 0,125 0,367 0,714 facilities -> satisfaction 0,284 0,288 0,120 2,373 0,018 facilities -> purchase 0,305 0,315 0,135 2,258 0,025 the value discriminant validity, ave and composite reliability, and path coefficient on fornell-larcker are shown in tables 4 – 6. table 6 shows path coefficients is also known as significance and a measure of strength. this figure is used to interpret the importance and strength of relationships between concepts. this table has a range of path factor values from – 0.05 to + 0.05 for path coefficients. value greater than 0.05 is considered a negative relationship, while a positive relationship is a value less than https://doi.org/10.34010/injiiscom.v4i1.9414 rudini et al., combination of technology acceptance model and decision-making … | 28 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 0.05, which increases the strength of the relationship. table 6 gives a detailed description of the table of path coefficients: h1a. evaluation of alternatives has an effect on trust. h1b. evaluation of alternatives has a significant effect on satisfaction. h1c. evaluation of alternatives has no effect on purchasing. h2a. usability has a significant effect on satisfaction. h2b. usability has no effect on satisfaction. h2c. usability has no effect on purchases. h3a. convenience has a significant and positive effect on purchases. h1b. evaluation of alternatives has a significant effect on satisfaction h2c. usability has no effect on purchases h3a. convenience has a significant and positive effect on purchasing h4a. trust has no effect on satisfaction h4b. trust has no effect on purchases h4c. satisfaction affects purchases h5a. seeking information has no effect on trust h5b. seeking information has no effect on satisfaction. 3.1. teory acceptance model (tam) this model was originally created by davis and has become one of the most widely used models to explain how users receive new technologies. this model was developed from the theory of reasoned action and provides a basis for identifying how external variables such as beliefs, attitudes, and intentions influence the acceptance of new technologies (see fig. 1) (dennis et al., 2010). perceived usefulness shows how far individuals will believe that utilizing technology improves the quality of their work (moe & fader, 2004). 3.2. perceived easy easy of use means that individuals believe that using information technology systems will not cause problems or require too much effort when used (hernández et al., 2011). 3.3. dmp (decision making process) the purchase decision process used in this study is related to the theory that the process is an orderly action and information search. table 7. r square adjust value r square r square adjusted belief 0.254 0.226 satisfaction 0.443 0.417 purchase 0.105 0.062 https://doi.org/10.34010/injiiscom.v4i1.9414 29 | international journal of informatics information system and computer engineering 4(1) (2023) 23-30 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 fig. 1. smart pls use tam & dma model this process includes searching for and gathering information about relevant products or services. it gives you a wide range of service products worth buying (cai et al., 2023). 3.4. evaluation of alternative involves evaluating and comparing different products or services worth buying (deananda et al., 2020). 4. conclusion customer satisfaction is influenced by the trust process. this research shows that customer attitudes and behaviors when shopping online are influenced by the process of customer trust in online stores. from this it follows that trust in online stores have a significant impact on customer attitudes and behavior. in order for the online shop business to succeed optimally, the goal is to maintain customer trust well and protect msme business actors who use information technology to further optimize sales. references bhatti, a., akram, h., basit, h. m., khan, a. u., mahwish, s., naqvi, r., & bilal, m. (2020). e-commerce trends during covid-19 pandemic. international journal of future generation communication and networking, 13(2), 1449–1452. cai, x., cebollada, j., & cortiñas, m. (2023). impact of sellerand buyer-created content on product sales in the electronic commerce platform: the role of informativeness, readability, multimedia richness, and extreme valence. journal of retailing and consumer services, 70 (september 2022). deananda, a., budiastuti, p., & muid, d. (2020). faktor-faktor pengaruh minat penggunaan sistem informasi akuntansi berbasis e-commerce pada aplikasi shopee dengan menggunakan technology acceptance model (tam). diponegoro journal accounting, 9(4), 1–10 https://doi.org/10.34010/injiiscom.v4i1.9414 rudini et al., combination of technology acceptance model and decision-making … | 30 doi: https://doi.org/10.34010/injiiscom.v4i1.9414 p-issn2810-0670 e-issn2775-5584 dennis, c., morgan, a., wright, l. t., & jayawardhena, c. (2010). the influences of social e-shopping in enhancing young women’s online shopping behaviour. journal of customer behaviour, 9(2), 151–174. fedorko, i., bacik, r., & gavurova, b. (2018). technology acceptance model in e commerce segment. management and marketing, 13(4), 1242–1256. haryanti, t., & subriadi, a. p. (2020). factors and theories for e-commerce adoption: a literature review. in international journal of electronic commerce studies (vol. 11, issue 2, pp. 87–105). academy of taiwan information systems research. hernández, b., jiménez, j., & martín, m. j. (2011). age, gender and income: do they really moderate online shopping behaviour? online informatio n review, 35(1), 113–133. kim, r. y. (2020). the impact of covid-19 on consumers: preparing for digital sales. ieee engineering management review, 48(3), 212–218. https://doi.org/10.1109/emr.2020.2990115 moe, w. w., & fader, p. s. (2004). dynamic conversion behavior at e-commerce sites. management science, 50(3), 326–335. https://doi.org/10.1287/mnsc.1040.0153 nagy, s., & hajdú, n. (2021). consumer acceptance of the use of artificial intelligence in online shopping: evidence from hungary.amfiteatru economic, 23(56), 1– 1. petcharat, t., & leelasantitham, a. (2021). a retentive consumer behavior assessment model of the online purchase decision-making process. heliyon, 7(10) riantini, r. e. (n.d.). adoption of e-commerce online to offline with technology acceptance model (tam) approach. rose, k., eldridge, s., & chapin, l. (2015). the internet of things (iot): an overview. int. journal of engineering research and applications, 5(12), 71–82. setyowati, w., widayanti, r., & supriyanti, d. (2021). implementation of e-business information system in indonesia: prospects and challenges. international journal of cyber and it service management, 1(2), 180–188. sheth, j. (2020). impact of covid-19 on consumer behavior: will the old habits return or die? journal of business research, 117, 280–283. https://doi.org/10.1016/j.jbusres.2020.05.059 wei, y., wang, c., zhu, s., xue, h., & chen, f. (2018). online purchase intention of fruits: antecedents in an integrated model based on technology acceptance model and perceived risk theory. frontiers in psychology, 9(aug). https://doi.org/10.34010/injiiscom.v4i1.9414 225 | international journal of informatics information system and computer engineering 3(2) (2022) 231-240 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 coffee tree detection using convolutional neural network ahmed abdulsalam naji hasan department of computer science, faculty of applied science taiz university, republic of yemen corresponding email: aaftername@gmail.com a b s t r a c t s a r t i c l e i n f o identifying plants is an important field in the environment because of their roles in the continuation of human existence. finding a plant by using the traditional methods such as looking at its physical properties is a burdensome task. thus, several computational-based methods have been introduced for detecting trees. in this study we constructed the coffee tree dataset due there is no publicly available coffee tree dataset for detection and classification of the coffee tree in orchard environments for what this tree has a role in health, industrial and agricultural fields, and raising the wheel of economic development. many machine learning algorithms have been used to detect and classify trees which resulted in reliable results. in this study, we presented a deep learning-based approach, in particular a convolutional neural network, for coffee tree detection and classification. the current study focused on providing a dataset for the detection and classification of coffee trees and improving the efficiency of the algorithm used in the detection and classification model. this study achieved the best results, the proposed system achieved an accuracy of 0.97%. article history: submitted/received 05 may 2022 first revised 01 jul 2022 accepted 19 sept 2022 available online 21 oct 2022 publication date 01 dec 2022 aug 2018 __________________ keywords: coffee tree dataset, almawasit, wadi balabil, al-ghayl, joreenat, al-zeila international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 231-240 https://doi.org/10.34010/injiiscom.v3i2.9504 ahmed abdulsalam naji hasan. coffee tree detection using convolutional…| 232 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 1. introduction 1.1. overview of the study agriculture is extremely important to continue human existence. it remains a major driver of many economies around the world, particularly in developing and underdeveloped economies. there is a growing demand for food and cash crops. machine learning techniques have been widely used in the control and management of agricultural crops, and the identification of types of agricultural crops. several machine learning algorithms have been used in the agricultural field since the beginning of the twentieth century. for example, convolutional neural networks (cnn) have been used to identify trees, it still needs need a standard data set designed to achieve high efficiency, by taking full advantage of from different assembly devices. to help identify and detect different crop trees. coffee is one of these important crops, driving forces of the economy in various countries of the world (clarence et al., 2003). for the importance of coffee 1 october 2015 world coffee daybasis of the economy of many of them, especially in the yemeni environment. thousands of families depend on the coffee crop to increase their annual income. approximately, one million people are working in this field starting from its cultivation until its exportation. it can be considered the main commodity that yemen exports to the world after oil. with the coffee crop, yemen has recorded a distinguished presence at the global level since the beginning of the sixth century as the first source of coffee (amamo, a. a.,2014). yemen is the only country in the world where the coffee tree is grown under climatic and environmental conditions that are not similar to other regions of the world. yemeni coffee, which is called arabica coffee, is considered the most famous and the most expensive coffee in the world (incomes et al., 2005). in the field of health, many studies have shown that coffee extracted from this tree is one of the substances that contain a great benefit to the body, as it treats many diseases such as type 2 diabetes colon and prostate cancer (kolb et al., 2020). and liver diseases (nieber et al., 2017). coffee also helps to increase focus and endurance (karayigit et al., 2021). it protects the liver from cirrhosis, as it reduces liver enzymes and prevents liver cancer, coffee is a rich source of antioxidants that protect people from tooth loss (priya et al., 2020). and gum disease (nagpal et al., 2014). thus, people have become more interested in planting coffee trees (alzaidi et al., 2016). this study aims to detect and classify the coffee tree using a convolutional neural network to encourage farmers to plant coffee trees. it will assist in developing the coffee tree in different countries, knowing the taxes of coffee trees, inventory, the farm management plan, and increasing the yield. detection and classification of the coffee tree are useful in making smart systems to classify many other important trees, especially those used in the field of medicine. 1.2. literature review computer vision has become very important in the knowledge of agricultural cash crops, and it is an easy and inexpensive way compared with the traditional methods (chandra et al., https://doi.org/10.34010/injiiscom.v3i2.9504 233 | international journal of informatics information system and computer engineering 3(2) (2022) 231-240 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 2020). several works have been introduced for detecting trees based on computer vision techniques. for example, zortea et al., presented a method for detecting citrus trees at highdensity orchards from images captured by unmanned aerial vehicles (uavs). they used the cnn method depicted in fig. 1 which provided good results ((zortea et al., 2018). fan et al., proposed a new algorithm based on deep neural networks, as shown in fig 2., to detect tobacco plants in images captured by uavs. the results showed that the proposed algorithm performs well (fan et al., 2018). wu et al., present an extracting apple tree crown information from remote imagery using (uavs).by using a faster r-cnn object detector. an automatically detect and segment individual trees and measure the crown width, perimeter, crown projection area of apple trees. the results were close to the manual delineation and this technique can bey used to detect and count apple an overall accuracy of 0.97%, estimate crown parameter with an overall accuracy exceeding 0.92% (wu et al., 2020). sun et al., presented image dataset collected by mobile phone in natural scene. which contains 10,000 images of 100 ornamental plant species in beijing forestry university campus. a 26-layer deep learning model consisting of 8 residual building blocks is designed for largescale plant classification in natural environment. the proposed model achieves a recognition rate of 0.91% on the bjfu100 dataset, demonstrating that deep learning is a promising technology for smart forestry (sun et al., 2017). santana et al., developed an algorithm for automatic counting of coffee plants and to determine the best age to carry out monitoring of plants using remotely piloted aircraft (rpa) images, it presented 96.8% accuracy with images without spectral treatment (santana et al., 2023). zheng et al., presented the cropdeep agricultural dataset. cropdeep species classification and detection dataset, consisting of 31,147 images with over 49,000 annotated instances from 31 different classes images were collected with different cameras and equipment in greenhouses, captured in a wide variety of situations. results show that current deep-learning-based methods achieve well performance in classification accuracy over 0.99% (zheng et al., 2019). diez et al., have published a review focused on studies that use dl and rgb images gathered by uavs to solve practical forestry research problems. the review discussed three main forestry problems including (1) individual tree detection, (2) tree species classification, and (3) forest anomaly detection (forest fires and insect infestation) this study useful for researchers that want to start working in this area (diez et al.,2021). gurumurthy et al., they presented a method for semantic segmentation of mango trees in high resolution aerial imagery, and, a novel method for individual crown detection of mango trees using segmentation output. results obtained demonstrate the robustness of the proposed methods despite variations in factors such as scale, occlusion, lighting conditions and surrounding vegetation (gurumurthy et al., 2019). https://doi.org/10.34010/injiiscom.v3i2.9504 ahmed abdulsalam naji hasan. coffee tree detection using convolutional…| 234 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 2. method in this chapter, we presented dataset of coffee tree, due to the lack of publicly available datasets to detect and classify the coffee tree, and system to detect and classify the coffee tree based on cnn deep learning algorithm for detection and classification of the coffee tree among more than one type of trees. as show in fig. 1. we briefly explain the main steps of the proposed method used to construct the dataset and model based on cnn technique (sk et al., 2021). the main phases are illustrated in the following table 1. table 1. proposed system phases dataset acquisition images preprocessing cnn classifier experiments testing of model 2.1. dataset acquisition we have constructed our own dataset from 417 coffee images and 493 other trees due to the lack of publicly available datasets to detect and classify the coffee tree, the images used in this work were collected from different regions of the yemeni environment. coffee trees and other trees were collected from the city of taiz, from al-mawasit department, from several orchards from the region of bani hammad, from wadi balabil and alghayl, as well as from several valleys in the region of joreenat and al-zeila. we collected rgb images of coffee tree and other types of trees. we used a similar standard (rodrguez et al., 2020). which are captured directly in place by midrange phone samsung j1 to capture the images (3 megapixels resolution, focus of f/2,2, srgb color jpg format 1536 × 2560 and 1920 × 1920 pixels). the pictures were taken at different distances in different lighting conditions. other images obtained from google image search with different format. we collected and stored them in dataset called (coffee tree dataset). we collected and stored them in dataset called (coffee tree dataset). combined in these ways in order to remove complexity for researchers and agricultural engineers. as shown in fig. 2. fig. 1. proposed system. https://doi.org/10.34010/injiiscom.v3i2.9504 235 | international journal of informatics information system and computer engineering 3(2) (2022) 231-240 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 fig. 2. coffee tree dataset. 2.2. images pre-processing the collected images of coffee trees and other trees used in this study are of different sizes, as shown in fig 3. thus, the first step we divided the data as follows: 60% for training, 20% for cross validation, and 20% for testing. when training the network, the input picture size is diversified, so that the network has better generalization (zheng et al., 2019). all of these transformations are contained within the imagedatagenerator. this function takes an image as input it. then, it uses a set of transformations such as increasing or decreasing brightness, flipping the image vertically or horizontally, rotating the image, shifting pixels. we processed these images, we used the rotation range=4, validation split = 0.20, rescale=1/255, width shift range=0.5, height shift range=0.5,shear range=0.10, zoom range=0.10,horizontal flip=true and fill mode=”nearest”. the second step we converted all of the images to the size of (100 ×100 × 3 pixels). (see fig 4). https://doi.org/10.34010/injiiscom.v3i2.9504 ahmed abdulsalam naji hasan. coffee tree detection using convolutional…| 236 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 fig. 3 images before preprocessing fig. 4 images after preprocessing 2.3. cnn classifier a deep convolutional neural network has become the dominating approach for image classification. year after year, various new architectures have been proposed. however, when choosing a deep architecture to solve a realistic problem, some aspects should be taken into consideration such as the type or number of layers, as a higher number of parameters increases the complexity of the system and directly influences the memory computation, speed and results of the system. although with specific characteristics according to realistic applications, a deep-learning network has the same goal to increase accuracy while reducing operation time and computational complexity. therefore, this study selected modern deep learning architecture. a model is proposed to detect and classify coffee tree among different trees. trees grading by human is inefficient, labor intensive (see fig. 4). from the equation shown in fig. 5, we used "adam" optimizer, we trained the model. during the training process, the model stopped at epoch=8. we noticed the model achieved better speed and good results. it achieved an accuracy 0.97% and an error rate of 0.04% is small, although the images were complex. https://doi.org/10.34010/injiiscom.v3i2.9504 237 | international journal of informatics information system and computer engineering 3(2) (2022) 231-240 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 fig. 5 equation for cnn classifier 3. testing of model we used 20% from coffee dataset a crossvalidation as an initial stage to evaluation the performance of the model and adjust some parameters after completing the training process for the model and to get good results. we tested the model, we used 20% from dataset. the model achieved an accuracy 0.75%, error rate 0.17% good results as long as the data is few, new and contains different characteristics compared to the training sample. we used a new data sample that had not been previously trained to test the model. fig. 4 cnn classifier after finishing we found that the model achieved better results. we also made predictions for two images one of the coffee tree images and one's unknown image, the images contain noise when looked at the images cannot be classified in a human way the two images have been uploaded to the model then we looked at the results we found the model correctly classified the images. https://doi.org/10.34010/injiiscom.v3i2.9504 ahmed abdulsalam naji hasan. coffee tree detection using convolutional…| 238 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 4. results through experiences we believe these results are satisfactory when visualized data with similar characteristics are considered, images inputs which were more contain noise. as well as the image taken of the coffee tree adjacent to other trees. despite these limitations, we achieved excellent results. we achieved 0.97% accuracy, precision 0.98%, recall 0.95% and the error rate very low 0.08%. 5. conclusions in this study, we proposed a new and fast technique to detect and classify the coffee tree. this study provided a dataset of coffee trees, this study proved that using the machine learning algorithm, the convolutional neural network (cnn) is able to detect and classify the coffee tree from images. convolutional layers can extract different abstract level features for a classification. we got an average accuracy 0.97%, an average error 0.04%. these results which achieved by the cnn algorithm are the best is very close to the features of manual measurement and visual inspection. this study can be relied upon instead of the traditional methods in detecting and classifying the coffee tree as well as other trees. references al-zaidi, a. a., baig, m. b., shalaby, m. y., & hazber, a. (2016). level of knowledge and its application by coffee farmers in the udeen area, governorate of ibb – republic of yemen. the journal of animal and plant sciences, 26, 1797-1804. amamo, a. a. (2014). coffee production and marketing in ethiopia. eur j bus manag, 6(37), 109-22essien, e. r., atasie, v. n., okeafor, a. o., & nwude, d. o. (2020). biogenic synthesis of magnesium oxide nanoparticles using manihot esculenta (crantz) leaf extract. international nano letters, 10(1), pp. 43-48. chandra, a. l., desai, s. v., guo, w., & balasubramanian, v. n. (2020). computer vision with deep learning for plant phenotyping in agriculture: a survey. arxiv preprint arxiv:2006.11391. clarence-smith, w. g., & topik, s. (eds.). (2003). the global coffee economy in africa, asia, and latin america, 1500–1989. cambridge university press. diez, y., kentsch, s., fukuda, m., caceres, m. l. l., moritake, k., & cabezas, m. (2021). deep learning in forestry using uav-acquired rgb data: a practical review. remote sensing,13(14), 2837. fan, z., lu, j., gong, m., xie, h., & goodman, e. d. (2018). automatic tobacco plant detection in uav images via deep neural networks. ieee journal of selected topics in applied earth observations and remote sensing,11(3), 876-887. https://doi.org/10.34010/injiiscom.v3i2.9504 239 | international journal of informatics information system and computer engineering 3(2) (2022) 231-240 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 gurumurthy, v. a., kestur, r., & narasipura, o. (2019). mango tree net–a fully convolutional network for semantic segmentation and individual crown detection of mango trees.arxiv preprint arxiv:1907.06915. incomes, s. i., & trade, e. (2005). moving yemen coffee forward. karayigit, r., naderi, a., akca, f., cruz, c. j. g. d., sarshin, a., yasli, b. c., ... & kaviani, m. (2021). effects of different doses of caffeinated coffee on muscular endurance, cognitive performance, and cardiac autonomic modulation in caffeine naive female athletes. nutrients, 13(1),2. kolb, h., kempf, k., & martin, s. (2020). health effects of coffee: mechanism unraveled?. nutrients, 12(6), 1842. nagpal, i. (2014). can milk, coffee and tea prevent dental caries?. international journal, 1(4), 129. nieber, k. (2017). the impact of coffee on health. planta medica, 83(16), 1256-1263. priya, s. l., jagannathan, r., balaji, t. m., varadarajan, s., venkatakrishnan, c., rajendran, s., ...\& devi, s. (2020). resveratrol and green coffee extract gel as anticaries agent. indian j. res. pharm. biotechnol, 8, 15-21. rodrguez, j. p., corrales, d. c., aubertot, j. n., & corrales, j. c. (2020). a computer vision system for automatic cherry beans detection on coffee trees.pattern recognition letters,136, 142-153. santana, l. s., santos, g. h. r. d., bento, n. l., faria, r. d. o. (2023). identification and counting of coffee trees based on convolutional neural network applied to rgb images obtained by rpa. sustainability, 15(1), 820. sk, rakibul, and ankita wadhawan. ”identification of plants using deep learning: a.” (2021). sun, y., liu, y., wang, g., & zhang, h. (2017). deep learning for plant identification in natural environment. computational intelligence and neuroscience, 2017. wu, j., yang, g., yang, h., zhu, y., li, z., lei, l., & zhao, c. (2020). extracting apple tree crown information from remote imagery using deep learning. computers and electronics in agriculture,174, 105504. zheng, y. y., kong, j. l., jin, x. b., wang, x. y., su, t. l., & zuo, m. (2019). cropdeep: the crop vision dataset for deep-learning-based classification and detection in precision agriculture. sensors, 19(5), 1058. https://doi.org/10.34010/injiiscom.v3i2.9504 ahmed abdulsalam naji hasan. coffee tree detection using convolutional…| 240 doi: https://doi.org/10.34010/injiiscom.v3i2.9504 p-issn 2810-0670 e-issn 2775-5584 zheng, y. y., kong, j. l., jin, x. b., wang, x. y., su, t. l., & zuo, m. (2019). cropdeep: the crop vision dataset for deep-learning-based classification and detection in precision agriculture. sensors, 19(5), 1058. zortea, m., macedo, m. m., mattos, a. b., ruga, b. c., & gemignani, b. h. (2018, october). automatic citrus tree detection from uav images based on convolutional neural networks. in2018 31th sibgrapi conference on graphics, patterns and images (sibgrapi)(vol. 11). https://doi.org/10.34010/injiiscom.v3i2.9504 1 gis-based urban village regional fire risk assessment and mapping yonathan andri hermawan *, lia warlina*, masnizah mohd** *departemen perencanaan wilayah dan kota, fakultas teknik dan ilmu komputer universitas komputer indonesia, jl. dipati ukur 102-116, bandung, 40132, indonesia **faculty of information science and technology, universiti kebangsaan malaysia 43600 bangi, selangor, malaysia *corresponding email:lia.warlina@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o fires in residential areas are one of the threats out of 13 disasters in indonesia. fires are disasters based on their causes, classified as disasters caused by human negligence. this research aims to identify residential fire incidents, assess fire risk levels, and map the risk level. we used the geographic information system (gis) analysis approach and direct observation of the study area. the research location was in the tamansari subdistrict in bandung city. the subdistrict of tamansari consists of 20 neighborhood units (rukun warga/ rw) with 22,995 people and 6,598 households. we conducted a field survey from december 2019 to march 2020. we used a spatial approach to analyze fire risk in this residential area by using gis to map urban-village regional fire incidents and assess the risk level. there were four fire hazard variables: population density, building density, building quality, road class. on the other hand, vulnerability variables are based on the community's social parameters: population density, percentage of old age and children under five, people with disabilities, and the population's sex ratio. the hazard and vulnerability maps overlay showed three neighborhood units (rukun warga/ rw) with a high risk of fire, eight rws with a moderate risk of residential fires, and nine rws with a low risk of residential fires. the areas with low-risk categories must remain vigilant because the width of the roads in these areas is relatively narrow. article history: ____________________ keyword: geographic information system (gis), urban village, fire, risk assessment international journal of informatics, information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics, information system and computer engineering 2 (2) (2021) 31-43 received 16 nov 2021 revised 20 nov 2021 accepted 25 nov 2021 available online 26 dec 2021 yonathan et al. gis-based urban village regional fire risk assessment and mapping|32 1. introduction fire is a disaster that, based on the causes of its occurrence, is classified as a natural disaster or a non-natural disaster caused by human negligence (man-made disaster). natural factors that cause fire disasters are lightning, earthquakes, volcanic eruptions, drought, and many others, while human factors are gas leaks, electrical short circuits, low construction security systems, and others (granda & ferreira, 2019; kumar, jana, & ramamritham, 2020; zhang, yao, & silanowicka, 2018). a fire in an area causes economic loss; therefore, it needs fire disaster management. urban fire disaster management research is carried out in many countries such as china, india, iran; with results showing that densely populated urban areas are at risk of fire (chan et al., 2018; kumar, ramamritham, & jana, 2019; navitas, 2014; waheed, 2014; zhang, yao, sila-nowicka, & jin, 2020). there are many methods for analyzing fire risk in urban areas. fire risk analysis for residential buildings in china uses cluster scenarios and applications (xin & huang, 2013). research in iran uses the fusion method of spatial information produced using unmanned aerial vehicles (uavs) and attribute data surveyed from 150 high-rise buildings (masoumi, van l.genderen, & maleki, 2019). meanwhile, in malmo, sweden, research on identifying the distribution of fires has made social stress one of the variables (guldåker & hallin, 2014). fire is one of the threats of 13 disasters in indonesia (badan nasional penanggulangan bencana, 2012). fig 1 shows the data on fire incidents in bandung city during 2019 (dinas kebakaran dan penanggulangan bencana kota bandung, 2020). the fire factor that often occurs in large cities such as bandung tends to be caused by human factors. the shape and planning of houses that are not regular make fire disasters often occur in bandung. the total population of bandung city in 2020 reached 2.444.160 people, with a population density of 14.61 thousand people per km2 (bpsstatistics of bandung municipality, 2021). fig 1. fire incidents in bandung city in 2019 (dinas kebakaran dan penanggulangan bencana, 2020) 10 12 14 16 13 26 16 43 39 51 19 13 0 10 20 30 40 50 60 yonathan et al. gis-based urban village regional fire risk assessment and mapping|33 this research will focus on one of the urban villages in bandung, a dense housing area located in kelurahan tamansari (tamansari sub-district). dense settlements that arise in the city of bandung, which is not accompanied by good supervision, have caused the spatial pattern of the residential areas to become irregular and very difficult to control. the housing density makes a fire more potential because of a poor water pipe system, very narrow road, and low-quality building materials. this research aims to identify residential fire incidents, assess fire risk levels, and map the risk level. 2. method the location of the research project was in tamansari sub-district. the study area consists of 20 neighborhood units (rukun warga/ rw) with 22,995 people and 6,598 households. we conducted the research project from december 2019 to march 2020. by using geographic information system (gis) analysis and direct observation of the study area. the data collection method is primary and secondary data. we used a spatial approach to analyze fire risk in this residential area. we applied gis to map urban-village regional fire incidents and assess the risk level. two main variables in this study to calculate fire risk are hazard and vulnerability. the level of disaster risk, using a formula, namely r = h x v/c (where: r = risk; h = hazard; v = vulnerability; c= capacity) (badan nasional penanggulangan bencana, 2012). capacity is the ability of regions and communities to reduce threats and losses due to disasters. this research does not include capacity in spatial analysis. there were four fire hazard variables, namely: population density, building density, building quality, road class (table 1). on the other hand, vulnerability variables are based on the community's social parameters: population density, percentage of elderly and toddlers, people with disabilities, and the population's sex ratio (table 2). table 1. fire hazard variables (badan nasional penanggulangan bencana, 2012) variables level of hazard weight population density < 150 persons/ ha (low) 1 150 200 persons/ ha (moderate) 2 >200 persons/ ha (high) 3 building density <40 units/ ha (low) 1 40-80 units/ ha (moderate) 2 >80 units/ha (high) 3 building quality <5% (low) 3 5 – 15 % (moderate) 2 > 15% (high) 1 road density >105 m/ha (high) 3 75-105 m/ha (modertae) 2 <75 m/ha (low) 1 yonathan et al. gis-based urban village regional fire risk assessment and mapping|34 table 2. fire vulnerability variables variables level of vulnerability weight population density < 150 persons/ ha (low) 1 150 200 persons/ ha (moderate) 2 >200 persons/ ha (high) 3 number of elderly and todler <20 % (low) 1 20-40 % (moderate) 2 >40% (high 3 number of disabled residents <20 % (low) 1 20-40 % (moderate) 2 >40% (high) 3 population sex ratio 92,38 – 98,88 (low) 1 98,89 – 105,39 (moderate) 2 105,4 – 111,9 (high) 3 3. results and discussion 3.1. distribution of fire locations in tamansari sub-district tamansari sub-district has a dense population density of up to 228 people/hectare due to urbanization. therefore, the need for land is very high, but the availability of land is insufficient, causing dense settlements in urban areas, which triggers a fire disaster that will occur due to the quality of the building and the unstable condition of the building material. based on data from the bandung city fire and disaster management service, there were fire incidents in the tamansari sub-district in 2015 – 2018. the fires in the tamansari subdistrict in 2015-2020 occurred at several points, and those were in rws of 1, 4, 9, 10, 11,15, and 20, period (fig 2). fig 3 shows one of the fire incidents in tamansari subdistrict. fig 2. fire incidents in taman sari sub-district in 2015-2020 yonathan et al. gis-based urban village regional fire risk assessment and mapping|35 fig 3. fire incident in tamansari subdistrict 3.2. fire hazard level the level of fire hazard was obtained from four fire hazard factors using the overlay method. table 3 shows the total weight of parameters is 11 (the maximum is 12). table 4 shows five rws with low levels, four rws with moderate levels, and 11 rws with a high fire hazard level. based on the residential fire hazard map, it appears that the tamansari village is dominated by a high fire hazard (fig 4). in contrast, the area with a low level of hazard is the trade and service area. the physical condition of the area is an essential factor in the fire hazard in urban village settlements, as studied in the bandung (permana, susanti, & wijaya, 2019) and surabaya city areas (navitas, 2014). figs 5 and 6 show the physical condition of the urban village in tamansari subdistrict. table 3. scores of fire hazard in tamansari subdistrict parameter fire hazard weight population density >200 person per hectare 3 building density >80 units per hectare 3 building quality 5 – 15 % 2 road density >105 meters per hectare 3 total 11 yonathan et al. gis-based urban village regional fire risk assessment and mapping|36 table 4. level and areas of fire hazard in tamansari subdistrict neighborhood units (rukun warga/ rw) area (hectares) level of hazard 1 7.7 low 2 6.7 low 3 17.2 low 4 1.9 high 5 1.7 high 6 1.8 high 7 4.8 moderate 8 3.4 low 9 2.3 moderate 10 5.2 low 11 5.3 high 12 4.5 high 13 2.6 high 14 1.7 high 15 7.3 high 16 3.8 high 17 7.9 moderate 18 7.4 high 19 4.1 moderate 20 4.6 high fig 4. map of fire hazard in tamansari subdistrict yonathan et al. gis-based urban village regional fire risk assessment and mapping|37 fig 5. dense settlement in urban village (tamansari subdistrict) fig 6. an example of low-quality building materials in tamansari subdistrict 3.3. fire vulnerability map vulnerability is a community or social condition that leads to or causes an inability to face the threat of disaster (trainor et al., 2009). the vulnerability parameter used in this study is the social vulnerability parameter. this social vulnerability is described through the condition of the population, which includes the sex ratio, disabled population, dependent age group of the elderly and infants, and population density where these factors can cause them to be in a vulnerable condition (sutanti, tjahjono, & syaufina, 2020; y. zhang, 2013). yonathan et al. gis-based urban village regional fire risk assessment and mapping|38 table 5 shows the results of the combined analysis of the four vulnerability factors above are further classified into three classes, namely low, medium, and high so that the level of fire vulnerability in the tamansari subdistrict has a score of 8. this result can be categorized into moderate residential fire vulnerability. table 6 shows the level and area of fire vulnerability in the tamansari subdistrict. there are nine rws with low levels, seven rws with moderate levels, and four rws with a high level of fire vulnerability. fig 7 shows the fire vulnerability of the urban village (tamansari subdistrict), caused by the social conditions of the people in the residential area. based on social conditions, the vulnerability level is dominated by rws with low levels. table 5. scores of fire vulnerability in tamansari subdistrict parameters fire vulnerability scores population density >200 persons/hectare (high) 3 number of elderly and toddler (%) 20-40 % (low) 1 number of disable resident (%) <20 % (low) 1 sex ration (%) 92.38 – 98.88 (low) 3 total 8 table 6. level and areas of fire vulnerability in tamansari subdistrict community units (rukun warga/ rw) area (hectares) level of vulnerability 1 7.7 low 2 6.7 low 3 17.2 low 4 1.9 moderate 5 1.7 moderate 6 1.8 low 7 4.8 low 8 3.4 low 9 2.3 high 10 5.2 low 11 5.3 moderate 12 4.5 moderate 13 2.6 high 14 1.7 low 15 7.3 high 16 3.8 high 17 7.9 moderate 18 7.4 moderate 19 4.1 low 20 4.6 moderate yonathan et al. gis-based urban village regional fire risk assessment and mapping|39 fig 7. map of fire vulnerability of tamansari subdistrict 3.4. fire risk table 7 and fig 8 describe the fire risk due to multiplying the hazard and vulnerability, showing three rws with high risk, eight rws with medium risk, and nine rws with low risk.fire risk shows that the area of the tamansari subdistrict is dominated by low risk. meanwhile, the low-risk area does not mean there is no potential for fire to occur. tamansari subdistrict has a narrow road width that makes it challenging to handle fires in residential areas, so it can be higher even though the risk is low. research on fire mitigation scenarios in dense settlements in sukahaji, bandung, showed that capacity optimization as a mitigation measure could be the main alternative in handling fire hazards in areas with medium-high population density. in addition, the early warning system is a crucial factor in mitigation efforts (sagala, adhitama, sianturi, & al faruq, 2016). to increase capacity in urban village housing, a proposed method by using existing resources for an emergency response include mosque loudspeakers, fire extinguishers, and preparing evacuation routes (pamungkas, rahmawati, larasati, rahadyan, & dito, 2017). yonathan et al. gis-based urban village regional fire risk assessment and mapping|40 table 7. scores of fire risk of tamansari subdistrict neighborhood unit (rw) the score of hazard (h) the score of vulnerability (v) risk(h x v) level of risk 1 1 1 1 low 2 1 1 1 low 3 1 1 1 low 4 3 2 6 moderate 5 3 2 6 moderate 6 2 1 2 low 7 2 1 2 low 8 2 1 2 low 9 2 3 6 moderate 10 1 1 1 low 11 3 2 6 moderate 12 3 2 6 moderate 13 3 3 9 high 14 3 1 3 low 15 3 3 9 high 16 3 3 9 high 17 2 2 4 moderate 18 3 2 6 moderate 19 2 1 2 low 20 3 2 6 moderate yonathan et al. gis-based urban village regional fire risk assessment and mapping|41 fig 8. map of fire risk of tamansari subdistrict 4. conclusion tamansari subdistrict is an urban village with dense settlements, which causes fire disasters to occur often. dense buildings and low building materials caused these incidents. based on the fire hazard and vulnerability analysis, tamansari subdistrict has a very high fire hazard level with a score of 11 of 12. the level of vulnerability based on social aspects shows that kelurahan tamansari has a low vulnerability value of 8. the fire risk map in the tamansari subdistrict is dominated by areas with a low risk of fires. the fire risk map shows three rws with high levels, eight rws with moderate levels, and nine rws with a low level of fire risk. 5. authors' note the authors declare that there is no conflict of interest regarding the publication of this article. the authors confirmed that the paper was free of plagiarism. 6. acknowledgment the authors would like to thank kelurahan tamansari for the data and information that is very useful for this research. yonathan et al. gis-based urban village regional fire risk assessment and mapping|42 7. references badan nasional penanggulangan bencana. (2012). peraturan kepala badan nasional penanggulangan bencana nomor 02 tahun 2012 tentang pedoman umum pengkajian risiko bencana. badan nasional penganggulangan bencana. bpsstatistics of bandung municipality. (2021). bandung municipality in figs 2021. chan, e. y. y., lam, h. c. y., chung, p. p. w., huang, z., yung, t. k. c., ling, k. w. k., … chiu, c. p. (2018). risk perception and knowledge in fire risk reduction in a dong minority rural village in china: a health-edrm education intervention study. international journal of disaster risk science, 9(3), 306–318. doi: 10.1007/s13753-018-0181-x dinas kebakaran dan penanggulangan bencana. (2020). data rekap kabakaran kota bandung [fire incident number]. bandung. retrieved from data.bandung.go.id/dataset/rekapitulasi-kejadian-kebakaran-di-kotabandung granda, s., & ferreira, t. m. (2019). assessing vulnerability and fire risk in old urban areas: application to the historical centre of guimarães. fire technology, 55(1), 105–127. doi: 10.1007/s10694-018-0778-z guldåker, n., & hallin, p.-o. (2014). spatio-temporal patterns of intentional fires, social stress and socio-economic determinants: a case study of malmö, sweden. fire safety journal, 70((2014)), 71–80. doi: 10.1016/j.firesaf.2014.08.015 kumar, v., jana, a., & ramamritham, k. (2020). a decision framework to assess urban fire vulnerability in cities of developing nations: empirical evidence from mumbai. geocarto international, 1–17. doi: 10.1080/10106049.2020.1723718 kumar, v., ramamritham, k., & jana, a. (2019). resource allocation for handling emergencies considering dynamic variations and urban spaces: fire fighting in mumbai. proceedings of the tenth international conference on information and communication technologies and development, 1–11. ahmedabad india: acm. doi: 10.1145/3287098.3287099 masoumi, z., van l.genderen, j., & maleki, j. (2019). fire risk assessment in dense urban areas using information fusion techniques. isprs international journal of geo-information, 8(12), 579. doi: 10.3390/ijgi8120579 navitas, p. (2014). improving resilience against urban fire hazards through environmental design in dense urban areas in surabaya, indonesia. procedia social and behavioral sciences, 135(2014), 178–183. doi: 10.1016/j.sbspro.2014.07.344 pamungkas, a., rahmawati, d., larasati, k. d., rahadyan, g. a., & dito, a. h. (2017). making a low risk kampong to urban fire. asian journal of applied sciences, 5(2), 367–375. doi: 10.24203/ajas.v5i2.4615 permana, a. y., susanti, i., & wijaya, k. (2019). kerentanan bahaya kebakaran di kawasan kampung kota. kasus: kawasan balubur tamansari kota bandung. jurnal arsitektur zonasi, 2(1), 32. doi: 10.17509/jaz.v2i1.15208 sagala, s. a. h., adhitama, p., sianturi, d. g., & al faruq, u. (2016). mitigation scenarios for residential fires in densely populated urban settlements in sukahaji village, bandung city. geoplanning: journal of geomatics and planning, 3(2), 147–160. doi: 10.14710/geoplanning.3.2.147-160 sutanti, n., tjahjono, b., & syaufina, l. (2020). analisis risiko bencana kebakaran di yonathan et al. gis-based urban village regional fire risk assessment and mapping|43 kecamatan tambora kota administrasi jakarta barat. tataloka, 22(2), 162– 174. doi: 10.14710/tataloka.22.2.162-174 trainor, s. f., calef, m., natcher, d., stuart chapin iii, f., mcguire, a. d., huntington, o., lovecraft, a. l. (2009). vulnerability and adaptation to climate-related fire impacts in rural and urban interior alaska. polar research, 28(1), 100–118. doi: 10.1111/j.1751-8369.2009.00101.x waheed, m. a. a. (2014). approach to fire-related disaster management in high density urban-area. procedia engineering, 77, 61–69. doi: 10.1016/j.proeng.2014.07.007 xin, j., & huang, c. (2013). fire risk analysis of residential buildings based on scenario clusters and its application in fire risk management. fire safety journal, 62((2013)), 72–78. doi: 10.1016/j.firesaf.2013.09.022 zhang, x., yao, j., & sila-nowicka, k. (2018). exploring spatiotemporal dynamics of urban fires: a case of nanjing, china. isprs international journal of geoinformation, 7(1), 7. doi: 10.3390/ijgi7010007 zhang, x., yao, j., sila-nowicka, k., & jin, y. (2020). urban fire dynamics and its association with urban growth: evidence from nanjing, china. isprs international journal of geo-information, 9(4), 218. doi: 10.3390/ijgi9040218 zhang, y. (2013). analysis on comprehensive risk assessment for urban fire: the case of haikou city. procedia engineering, 52, 618–623. doi: 10.1016/j.proeng.2013.02.195 41 | international journal of informatics information system and computer engineering 3(2) (2022) 41-49 new modern approach to predict users’ sentiment using cnn and blstm r. sathish kumar* * department of computer science and engineering manakula vinayagar institute of technology, kalitheerthalkuppam, puducherry, india. *corresponding email: sathishmail8@gmail.com a b s t r a c t s a r t i c l e i n f o in today’s world social network play a vital role and provides relevant information on user opinion. this paper presents emotional health monitoring system to detect stress and the user mood. depending on results the system will send happy, calm, relaxing or motivational messages to users with phycological disturbance. it also sends warning messages to authorized persons in case a depression disturbance is detected by monitoring system. this detection of sentence is performed through convolution neural network (cnn) and bi-directional long-term memory (blstm). this method reaches accuracy of 0.80 to detect depressed and stress users and also system consumes low memory, process and energy. we can do the future work of this project by also including the sarcastic sentences in the dataset. we can also predict the sarcastic data with the proposed algorithm. article history: received 18 dec 2022 revised 20 dec 2022 accepted 25 dec 2022 available online 26 dec 2022 aug 2018 __________________ keywords: sentimental analysis, recommendation system, deep learning, cnn, blstm, social networks. 1. introduction nowadays, the number of active social network users has grown drastically. this high number of users on social networks is mainly due to the increase of the number of mobile devices, such as smart phones and tablets. currently osn have become a universal means of opinion, expression, feelings, and they reflect the bad habits or wellness practices of each user. in recent years, the analysis of the messages posted on social networks have been used by many applications in the industry of healthcare informatics. at first, social media existed to help end users connect digitally with friends, colleagues, family members, and international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 41-49 r. sathish kumar. new modern approach to predict users’ sentiment using …| 42 like-minded individuals they might never have met in person. desktop access to bulletin board services such as compuserve and prodigy made it easier to grow free online communities without ever leaving the house. as social media companies grew their user bases into the hundreds of millions, the business applications of facebook, twitter, and other social platforms began to take shape. social media companies had access to some of the richest trackable user data ever conceived. another downside of many of the internet’s segmented communities is that users tend to be exposed only to information they are interested in and opinions they agree with. this lack of exposure to novel ideas and contrary opinions can create or reinforce a lack of understanding among people with different beliefs, and make political and social compromise more difficult to come by. 2. literature survey 2.1. the use of social media for communication social media takes on many different forms including magazines, internet forums, weblogs, social blogs, micro blogging, wikis, podcasts, photographs or pictures, video, rating and social bookmarking (glavan et al., 2016; who, 2016; zhang et al., 2018). with the world in the midst of a social media revolution, it is more than obvious that social media like face book, twitter, orkut, myspace, skype etc., are used extensively for the purpose of communication. 2.2. music recommendation system based on user’s sentiments extracted from social networks. this paper presents a music recommendation system based on a sentiment intensity metric, named enhanced sentiment metric (esm) that is the association of a lexicon-based sentiment metric with a correction factor based on the user's profile (rosa et al., 2018; al-qurishi et al., 2018; sathish kumar & pariselvam, 2012). this correction factor is discovered by means of subjective tests, conducted in a laboratory environment. based on the experimental results. 2.3. hunting suicide notes in web2.0 preliminary findings: y. p. huang, t. goh, and c. l. liew. this paper will explore the techniques used by other researchers in the process of identifying emotional content in unstructured data, and will make use of existing technologies to attempt to identify at-risk bloggers (huang et al., 2007; sathish et al., 2016; sathish et al., 2016; sathish et al., 2017). using a selection of real blog entries harvested from myspace.com, supplemented with artificial entries from our research, we test the accuracy of a simple algorithm for scoring the presence of certain key words and phrases in blog entries. despite the simplistic approach taken, the preliminary results of this study were very promising. 2.4. detecting stress based on social interactions in social networks: in this paper, we find that users stress state is closely related to that of his/her friends in social media, and we employ a large-scale dataset from real-world social platforms to systematically study the 43 | international journal of informatics information system and computer engineering 3(2) (2022) 41-49 correlation of users' stress states and social interactions (lin et al., 2017; sathish et al., 2019; thapliyal et al., 2017; berbano et al., 2017; sathish et al., 2019). we first define a set of stress-related textual, visual, and social attributes from various aspects, and then propose a novel hybrid model a factor graph model combined with convolution neural network to leverage tweet content and social interaction information for stress detection. 2.5. deep learning based document modeling for personality detection from text: the authors train a separate binary classifier, with identical architecture, based on a novel document modeling technique. namely, the classifier is implemented as a specially designed deep convolution neural network, with injection of the document-level marissa features, extracted directly from the text, into an inner layer (majumeder et al., 2017; xue et al, 2014; tsugawa et al., 2015; sathish et al., 2020; rodrigues et al., 2016). the first layers of the network treat each sentence of the text separately; then the sentences are aggregated into the document vector. filtering out emotionally neutral input sentences improved the performance. this method outperformed the state of the art for all five traits, and the implementation is freely available for research purposes. 3. existing system machine learning is that field of study that provides computers the aptitude to find out while not being expressly programmed. millilitre is one of the foremost exciting technologies that one would have ever come upon. because it is clear from the name, it provides the pc that creates it additional the same as humans the power to find out. machine learning is actively being employed these days, maybe in more places than one would expect (ma & hovy, 2016; lample et al., 2016; khodayar et al., 2017; guimaraes et al., 2017; araque et al., 2017). 3.1. types of machine learning: machine learning implementations are classified into 3 major classes, betting on the character of the training “signal” or “response” on the market to a learning system. 1. supervised learning 2. unsupervised learning 3. reinforcement learning 4. semi-supervised learning 1. supervised learning when associate degree algorithmic rule learns from example knowledge and associated target responses which will carry with it numeric values or string labels, like categories or tags, so as to later predict the proper response once displayed with new examples comes beneath the class of supervised learning. 2. unsupervised learning whereas once associate degree algorithmic rule learns from plain examples with none associated response, going away to the algorithmic rule to work out the info patterns on its own. this sort of algorithmic rule tends to reconstitute the info into one thing else, like new options that will represent a r. sathish kumar. new modern approach to predict users’ sentiment using …| 44 category or a brand-new series of uncorrelated values. 3. reinforcement learning. when you give the rule with examples that lack labels, as in unsupervised learning. however, you may accompany associate example with positive or feedback per the solution the rule proposes comes beneath the category of reinforcement learning, that's connected to applications that the rule ought to produce picks (so the merchandise is prescriptive. 4. semi-supervised learning. where associate degree incomplete coaching signal is given: a coaching set with some (often many) of the target outputs missing. there's a special case of this principle called transduction wherever the complete set of downside instances is understood as learning time, except that a part of the target area unit is missing (fig. 1). fig. 1. flow diagram – ml working a common example of an application of semi-supervised learning is a text document classifier. this is the type of situation where semi-supervised learning is ideal because it would be nearly impossible to find a large amount of labeled text documents. this is simply because it is not time efficient to have a person read through entire text documents just to assign it a simple classification. so, semi-supervised learning allows for the algorithm to learn from a small amount of labeled text documents while still classifying a large amount of unlabeled text documents in the training data (fig. 2). fig. 2. flow diagram – data splitting 3.2. limitations of existing work: • the existing system shows accuracy of only 60% • it uses only random forest algorithm to process the data. • its efficiency of processing the results is slow and so the results appear • its user interface is not friendly. • it shows error in some results. • it does not process all the data; it leaves some of the data. 5. proposed work 5.1. algorithms used 1) random forest random forest may well be a machine learning formula that belongs to the supervised learning technique. it is typically used for every classification and regression problem in cubic 45 | international journal of informatics information system and computer engineering 3(2) (2022) 41-49 centimeter. it’s supported the conception of ensemble learning. random forest is one of the best high-performance strategies widely applied in numerous industries due to its effectiveness. it can handle data very effectively, whether it is binary, continuous, or categorical. random forest is difficult to beat in terms of performance. of course, you can always discover a model that performs better, for example, neural networks. still, they take longer to construct and can handle a wide range of data types, including binary, category, and numerical. one of the finest aspects of the random forest is that it can accommodate missing values, making it an excellent solution for anyone who wants to create a model quickly and efficiently (fig. 3). fig. 3. architecture diagram – diagram random forest 2) blstm algorithm: blstm is an associate degree extension of ancient lstm. it will improve model performance on sequence classification issues. in issues wherever all time steps of the input sequence area unit offered, blstm train 2 rather than one lstm on the input sequence. lstm in its core preserves information from inputs that has already passed through it using the hidden state. unidirectional lstm only preserves information of the past because the only inputs it has seen are from the past. using bidirectional will run your inputs in two ways, one from past to future and one from future to past and what differs this approach from unidirectional is that in the lstm that runs backwards you preserve information from the future and using the two hidden states combined you are able in any point in time to preserve information from both past and future. 3) convolutional neural network (cnn): a convolution is the straightforward application of a filter to associate degree input that leads to activation. indicating the locations associate degree strength of a detected feature in an input, like a picture. the innovation of cnn is the ability to mechanically learn an outsized range of filters in parallel specific to a coaching dataset below the constraints of a particular prognostication modelling drawback, like image classification. the experimental analysis shows that the model gives the accuracy of 84%. 6. results and discussion 6.1. upload osn dataset using this module, we will upload dataset to application. user profile and user data: database built from the data captured from osns. messages: there is a database with 360 messages, 90 messages for each kind (relaxing, motivational, happy, or calm messages) r. sathish kumar. new modern approach to predict users’ sentiment using …| 46 to be suggested to the user by the recommendation engine. the users can previously choose one or two kinds of messages when they undergo a period of stress or depression. the messages were written by 3 specialists in psychology and validated by 3 other specialists. 6.2. generate train & test model from osn dataset: using this module, we will read all messages from dataset and build a train and test model by extracting features from dataset. depression or stress detection by machine learning: the sentences are extracted from osn and they are filtered by machine learning to detect depression or stress conditions. it is implemented in the emotional health monitoring system. 6.3. build cnn blstm-rnn model using softmax: using this module, we will build deep learning blstm model on dataset and then using test data we will calculate blstm prediction accuracy. the cnns have several different filters/kernels consisting of trainable parameters which can convolve on a given image spatially to detect features like edges and shapes. hence, they can successfully boil down a given image into a highly abstracted representation which is easy for predicting. 6.4. upload test message & predict sentiment & stress: using this module, we will upload test messages and then application will detect stress by applying blstm model on test data. in the proposed system, users’ personal information and context information is used. however, users do not always post this related information. in case users do not post personal information, standard information is used, such as sleep routine of 8 hours, no unhealthy habits, no preferences about work or study. it is important to note that in our tests only 5% of the users do not post this information. a traditional rs is also implemented, in which only the words searched by a person on osn are used to feed the system, forming a content-based rs. for the sake of simplicity, the traditional content-based rs will not be explained in this section (figs. 4-10). fig. 4. in fig. 4 i am uploading the dataset file which contains messages. fig. 5. in the above screen we can see the records for to test the prediction performance. 47 | international journal of informatics information system and computer engineering 3(2) (2022) 41-49 fig. 6. blstm model generated and the accuracy is shown as 83.94%. fig. 7. in the above screen we can see the iterations to generate prediction layers. fig. 8. the random forest prediction accuracy is 36% which is lower than proposed blstm accuracy. fig. 9. in the above screen we can see each message application detected and mark with stress or nonstress status. fig. 10. accuracy of both the algorithms. 7. conclusion various deep-learning techniques can be used for the prediction of sentimental analysis and recommendation. the challenge is to develop accurate and computationally efficient medical data classifiers. in this paper the model contains the emotional health monitoring system, which uses the deep learning model and the sentiment metric named esm2. the sentences are extracted from an osn and then emotional health monitoring system identifies which sentences present a stress or depression r. sathish kumar. new modern approach to predict users’ sentiment using …| 48 content using machine learning algorithms and the emotion of the sentence content. references al-qurishi, m., hossain, m. s., alrubaian, m., rahman, s. m. m., & alamri, a. (2017). leveraging analysis of user behavior to identify malicious activities in largescale social networks. ieee transactions on industrial informatics, 14(2), 799813. araque, o., corcuera-platas, i., sánchez-rada, j. f., & iglesias, c. a. (2017). enhancing deep learning sentiment analysis with ensemble techniques in social applications. expert systems with applications, 77, 236-246. berbano, a. e. u., pengson, h. n. v., razon, c. g. v., tungcul, k. c. g., & prado, s. v. (2017, september). classification of stress into emotional, mental, physical and no stress using electroencephalogram signal analysis. in 2017 ieee international conference on signal and image processing applications (icsipa) (pp. 11-14). ieee. glavan, i. r., mirica, a., & firtescu, b. n. (2016). the use of social media for communication in official statistics at european level. romanian statistical review, 64(4), 37-48. guimaraes, r. g., rosa, r. l., de gaetano, d., rodriguez, d. z., & bressan, g. (2017). age groups classification in social network using deep learning. ieee access, 5, 10805-10816. huang, y. p., goh, t., & liew, c. l. (2007, december). hunting suicide notes in web 2.0-preliminary findings. in ninth ieee international symposium on multimedia workshops (ismw 2007) (pp. 517-521). ieee. khodayar, m., kaynak, o., & khodayar, m. e. (2017). rough deep neural architecture for short-term wind speed forecasting. ieee transactions on industrial informatics, 13(6), 2770-2779. kumar, r. s., & pariselvam, s. (2012). formative impact of gauss markov mobility model on data availability in manet‖. asian journal of information technology, 11(3), 108-116. kumar, r. s., dhinesh, t., & kathirresh, v. consensus based algorithm to detecting malicious nodes in mobile adhoc network. international journal of engineering research & technology (ijert) vol, 6. kumar, r. s., koperundevi, s., & suganthi, s. (2016). enhanced trust based architecture in manet using aodv protocol to eliminate packet dropping attacks. international journal of engineering trends and technology, 34, 21-27. kumar, r., logeswari, r., devi, n., & bharathy, s. (2017). efficient clustering using ecatch algorithm to extend network lifetime in wireless sensor networks. int. j. eng. trends technol, 45, 476-481. 49 | international journal of informatics information system and computer engineering 3(2) (2022) 41-49 lample, g., ballesteros, m., subramanian, s., kawakami, k., & dyer, c. (2016). neural architectures for named entity recognition. arxiv preprint arxiv:1603.01360. lin, h., jia, j., qiu, j., zhang, y., shen, g., xie, l., ... & chua, t. s. (2017). detecting stress based on social interactions in social networks. ieee transactions on knowledge and data engineering, 29(9), 1820-1833. ma, x., & hovy, e. (2016). end-to-end sequence labeling via bi-directional lstm-cnnscrf. arxiv preprint arxiv:1603.01354. majumder, n., poria, s., gelbukh, a., & cambria, e. (2017). deep learning-based document modeling for personality detection from text. ieee intelligent systems, 32(2), 74-79. rodrigues, r. g., das dores, r. m., camilo-junior, c. g., & rosa, t. c. (2016). sentihealth-cancer: a sentiment analysis tool to help detecting mood of patients in online social networks. international journal of medical informatics, 85(1), 80-95. rosa, r. l., rodriguez, d. z., & bressan, g. (2015). music recommendation system based on user's sentiments extracted from social networks. ieee transactions on consumer electronics, 61(3), 359-367. sathish kumar r, abdulla m.g. (2019). “head gesture and voice control wheel chair system using signal processing”, asian journal of information technology, 18(18), issue 8, 1682-3915. sathish kumar r, girivarman r, parameshwaran s, sriram v. (2020) "stock price prediction using deep learning and sentimental analysis", international journal of emerging technologies and innovative research 7(5), 346-354. sathishkumar, r., kalaiarasan, k., prabhakaran, a., & aravind, m. (2019, march). detection of lung cancer using svm classifier and knn algorithm. in 2019 ieee international conference on system, computation, automation and networking (icscan) (pp. 1-7). ieee. thapliyal, h., khalus, v., & labrado, c. (2017). stress detection and management: a survey of wearable smart health devices. ieee consumer electronics magazine, 6(4), 64-69. tsugawa, s., kikuchi, y., kishino, f., nakajima, k., itoh, y., & ohsaki, h. (2015, april). recognizing depression from twitter activity. in proceedings of the 33rd annual acm conference on human factors in computing systems (pp. 3187-3196). world health organization. (2016). world health statistics 2016: monitoring health for the sdgs sustainable development goals. world health organization. xue, y., li, q., jin, l., feng, l., clifton, d. a., & clifford, g. d. (2014, april). detecting adolescent psychological pressures from micro-blog. in international conference on health information science (pp. 83-94). springer, cham. zhang, y., xu, c., li, h., yang, k., zhou, j., & lin, x. (2018). healthdep: an efficient and secure deduplication scheme for cloud-assisted ehealth systems. ieee transactions on industrial informatics, 14(9), 4101-4112. 89 | international journal of informatics information system and computer engineering 4(1) (2023) 89-100 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 development of an educational training game for ear sensitivity of intervals michael nagaku milenn salim, hanhan maulana* program studi teknik informatika, universitas komputer indonesia , indonesia *corresponding email: hanhan@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o this study aims to build a game to help beginner musicians in the process of training ear sensitivity to intervals without involving the teacher in the training process. the software development uses the multimedia development life cycle (mdlc) method. a software is built in the form of a game that can train the ear's sensitivity to intervals, share knowledge related to intervals, randomize questions and validate answers without involving the teacher in the learning process. the average respondents' answers related to the application were positive. it can be concluded that the development of the application can help users in training the ear sensitivity to intervals and increase the user's knowledge of music theory, especially related to intervals. article history: submitted/received 20 dec 2022 first revised 03 jan 2023 accepted 07 mar 2023 first available online 18 apr 2023 publication date 01 jun 2023 aug 2018 __________________ keywords: game, ear training, music, multimedia, intervals 1. introduction in the world of music, the pitch interval is the distance between one note and another (levine, m. 2011). interval is what builds music, when we listen to music, we listen to intervals (willis, g. 1998). a musician, both beginner and expert must know the pitch interval. the ability to identify the intervals of the notes heard is very helpful in musical activities. an ear that is sensitive to pitch intervals is a musician's most valuable asset (pavlik, p. i., et al., 2013) games are fun entertainment media that can be played to fill spare time. games can also be used as fun learning media, commonly called educational games (jubaedi, a. d., & putra, r. e. 2018) (arsenault, d. 2009). a fun learning process for what is learned can make the subject matter more interesting and facilitate the delivery of the subject matter international journal of informatics information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 4(1) (2023) 89-100 https://doi.org/10.34010/injiiscom.v4i1.9588 michael and hanhan. development of an educational training game for ear ,...| 90 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 (pradana, a. g. 2019). the use of educational games in learning can improve the quality and disruption in the learning process as well as train the user's ability to solve problems, find solutions, think quickly and improve (limin, s. 2022), (ardiningsih, d. 2019). a good musician has an ear that is sensitive to changes in pitch. beginner music generally knows the concept of solfge (do, re, mi, fa, sol, la, si, do) but does not yet have a good ear. the ear's sensitivity to pitch intervals can develop over time. development may be aided by ear or ear repair exercises (willis, g. 1998). the music learning process, in this case ear training, requires a teacher who can determine questions randomly and dynamically and correct errors during the learning process. to have someone who teaches can cost a lot of money, especially face-to-face learning during a pandemic can pose a danger of spreading covid-19 (santoso, a. m. h. 2022). the need to do social distancing complicates the process of learning music, especially in terms of ear training. by using educational games as learning media in terms of ear training, you can learn to automate the randomization of questions dynamically as well as the process of correcting errors in social distancing conditions. the use of educational games can also reduce the cost of learning music. several studies related to games for ear training or music theory have been carried out. in a study entitled "adventure game as learning media for introducing music interval and ear training to kids" discusses the use of games as an educational medium for musical intervals and ear training for children (rizqyawan, m. i., & hermawan, g. 2015). another study that discusses ear training with the title "development of interactive quiz games as a formative evaluation instrument in music theory course" builds games for student understanding in music theory courses (haditama, i., et al., 2016). several previous studies built desktop-based singleplayer games with 2d graphics. however, in studies, input from users is still in the form of a multiple-choice quiz. ear training exercises by singing or humming can help the ear hear better intervals. therefore, it requires input in the form of sound through a microphone so that the learning process can be more effective (willis, g. 1998). the software method that will be used is the multimedia development life cycle method or abbreviated as mdlc. mdlc was chosen as the development method because educational games are part of interactive multimedia so mdlc is the right method for developing this software (afrianto, i., & furqon, r. m. 2018). 2. method in this study, the methodology used includes three stages in fig. 1, with the following explanation: https://doi.org/10.34010/injiiscom.v4i1.9588 91 | international journal of informatics information system and computer engineering 4(1) (2023) 89-100 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 fig. 1. research methodology 2.1. data collection stage at this stage, data is collected that will be used in research. the data collection process begins with conducting a literature study related to research and technology in the field of music, especially research that integrates technology in the field of music. literature study was conducted to find out what research has been done, expand knowledge and widen the range of ideas. next is the observation stage, which will carry out the problem identification process from the results of these observations. 2.2. educational game development stage the approach used in the development of educational games is the multimedia development life cycle (mdlc). in this approach to the six stages that need to be carried out, the stages are (mursid, r. 2018), (laksamana, d.j., et al 2021): i) concept the first stage is the concept, in this study raised the concept of educational games with the theme of outer space for learning music theory and training ear sensitivity. ii) design the second stage is the stage where the design is carried out, both menu, character and storyboard designs. iii) material collecting the third stage is the stage for collecting material, the material collected is in the form of audio, images, and data. iv) assembly the fourth stage is the development stage where the components are built into a complete system. v) testing the fifth stage is the testing stage, the testing process will be carried out at this stage. vi) distribution the sixth stage is the distribution stage, in order to disseminate research on this educational game to be useful for the community. the activities of the mdlc approach can be seen in fig 2. https://doi.org/10.34010/injiiscom.v4i1.9588 michael and hanhan. development of an educational training game for ear ,...| 92 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 fig. 2. mdlc approach 2.2.1. concept the game that was built is a game that helps beginner musicians in the process of training ear sensitivity to tone intervals. the game built is a desktopbased training educational game with 2d (2 dimensional) graphics. the related description of the game built as follows: i) the game will play questions in the form of interval sounds with the same root in each question ii) players guess the sound by providing input in the form of the same interval sound through the microphone iii) the game will calculate the player's score on each level iv) the next level will be unlocked if the player reaches a minimum score of 75% v) each level has a different difficulty vi) the material to be taught and trained is part of the c major scale vii) game is a training which must be done repeatedly. 2.2.2. design the design stage is the stage of carrying out the specifications of the game that is built. the system design built is: i) storyline design a spaceship pilot assigned to collect precious space rocks. spaceship pilots must sail through the vast darkness of outer space in order to collect the precious stones needed by the main ship. players will play a spaceship character. ii) level design the levels are divided into seven sections. each level will practice different pitch intervals. players will start the game from level 1. the next level will be unlocked when the player has reached the minimum score. the minimum score that needs to be achieved to unlock the next level feature is 75%, 75% is obtained from the national completeness target (yusuf, m. 2019), (aldwell, e., et al., 2018). the list of tone intervals that are trained at each level can be seen in table 1. https://doi.org/10.34010/injiiscom.v4i1.9588 93 | international journal of informatics information system and computer engineering 4(1) (2023) 89-100 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 iii) storyboard design displays a series of notes vertically, spaceship characters, space rocks as a hint of tone intervals. the game will play a matter of tone intervals in the form of sound. if the player activates the hint feature, a stone will be displayed which is an indication of the tone that must be played. if the player does not activate the hint feature, the rocks become invisible. players will provide input in the form of sound through the microphone. the character of the aircraft will move up or down according to the input frequency of the tone from the user. the lowest “do” tone represents c3. the storyboard design can be seen in fig 3. table 1. intervals in each level level trained intervals 1 unison, major second 2 unison, major second, major third 3 unison, major second, major third, perfect fourth 4 unison, major second, major third, perfect fourth, perfect fifth 5 unison, major second, major third, perfect fourth, perfect fifth, major sixth 6 unison, major second, major third, perfect fourth, perfect fifth, major sixth, major seventh, 7 unison, major second, major third, perfect fourth, perfect fifth, major sixth, major seventh, perfect octave 8 ghundul-ghundul pacul song https://doi.org/10.34010/injiiscom.v4i1.9588 michael and hanhan. development of an educational training game for ear ,...| 94 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 fig 3. storyboard design iv) gameplay design this part is a gameplay where players will be trained to be sensitive to their ears when faced with obstacles. players will be represented as a spaceship that blocks the vast darkness of outer space. in this mode, the player must capture as many spaces precious stones as possible. players will set the spaceship's flight path with sound through the microphone. the movement of the spaceship will refer to the input tone (frequency) provided by the player. v) scoring design the score at each level will depend on how many intervals the player can guess correctly. the total predictable interval will be divided by the total number of questions at each level. the scoring calculation formula can be seen in formula (1). 𝑆 = 𝐵 𝑁 × 100% (1) with the following variables: s = score players at that level. b = number of correctly guessed intervals in the level. n = the total number of questions in the level. vi) character design players will play a spaceship character, and the questions played will appear as space aids if players use the hint feature. the character designs used in the game can be seen in table 2. 2.2.3. material collecting the material collecting stage is the stage of collecting materials or assets used in the process of developing educational games. the assets collected can be seen in table 3. https://doi.org/10.34010/injiiscom.v4i1.9588 95 | international journal of informatics information system and computer engineering 4(1) (2023) 89-100 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 table 2. character design no name figure information 1. pesawat luar angkasa in-game player character 2. bebatuan do-si characters for do to si tone hints with different colors table 3. assets used no name information 1. play button play button asset to enter select window 2. quit button asset quit button to leave the game 3. back button asset back button to return to the previous window 4. hint tidak aktif button inactive hint key asset indicates that the hint feature is not active 5. hint aktif button the active hint key asset indicates that the hint feature is active 6. score button score button asset to display the jscore jendela window 7. go button button asset to display the jgo jendela window 8. laser biru soal the limiting asset for segment markers of voice questions 9. laser kuning jawab limiting asset for voice input segment marker 10. laser hijau jawab benar limiting asset for voice input segment marker, voice input marker from correct player 11. laser merah jawab salah limiting asset for voice input segment marker, voice input marker from wrong player 12. tombol level 1-8 key asset level 1-8 to display the jlevel1-8 . window 13. papan penjelasan explanation board assets to display explanatory information on each level 2.2.4. assembly the assembly stage is the implementation stage or combining assets and designs into a complete game that is ready to use. the main menu interface, or the first display in the game can be seen in fig 4. figure 5 is a level selection interface with a hint feature tool. figure 6 is the interface that appears when the user has selected a level. the content level interface also displays the materials at that level. the levels that have been taught will be trained on the main game interface. the main game interface is shown in fig 7. https://doi.org/10.34010/injiiscom.v4i1.9588 michael and hanhan. development of an educational training game for ear ,...| 96 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 fig. 4. home menu interface fig. 5. level selects menu interface fig. 6. level material interface fig. 7. main game interface after the user completes the main game, the score obtained by the user will be saved by the system which can be viewed again on the score interface. the score interface can be seen in fig 8. fig. 8. score interface 2.2.5. testing the testing stage is the stage of testing the game that was built. the tests carried out are blackbox testing, musical instrument testing and beta testing. 2.2.6. distribution the distribution stage is the stage where the game will be uploaded to google drive which users will download later. 2.3. conclusion drawing stage the conclusion stage is the stage where conclusions will be drawn based on responses from users who have played the game that was built. 3. results and discussion the games that have been built will be tested using the blackbox method, then musical instrument testing and beta testing will be carried out which will involve users. the first test is blackbox testing which aims to find functional errors or bugs in the game. testing the main components in blackbox can be seen in table 4. https://doi.org/10.34010/injiiscom.v4i1.9588 97 | international journal of informatics information system and computer engineering 4(1) (2023) 89-100 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 the next test is the testing of musical instruments, this test is carried out to test whether the musical instruments used can be received well by the game that is built. if the sound can be processed properly, the spacecraft will move up or down following the given sound frequency. the results of testing musical instruments can be seen in table 5. the next test is beta testing, this test will involve users as respondents who will answer the questions that have been provided previously. the questions that have been provided can be seen in table 6. respondents will answer the questions in table 6 with multiple choice answers with their respective weights, the weight of each answer can be seen in table 7. the results of beta testing for games that have been built on previously prepared questions can be seen in table 8. table 4. blackbox test results no tested components testing scenario test results 1. home menu buttons on the start menu succeed 2. menu select level buttons on the level menu succeed selecting level 1 to level 2 succeed 3. level menu the buttons on the level menu succeed 4. gameplay move the plane with the microphone succeed answer the questions correctly succeed answering questions incorrectly succeed the buttons on the level menu succeed 5. score save game score succeed view game score succeed table 5. musical instrument test results musical instrument c3 d3 e3 f3 g3 a3 b3 c4 nylon stringed guitar ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ steel stringed guitar ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ guitalele ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ human voice ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ https://doi.org/10.34010/injiiscom.v4i1.9588 michael and hanhan. development of an educational training game for ear ,...| 98 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 table 6. beta testing questions no question 1 is by playing this game your knowledge of music theory related to tone intervals increases? 2 can the question of the interval of the tone played be clearly heard? 3 is the ear training provided according to your skill level? 4 can this game provide an understandable pitch interval theory? 5 do you feel comfortable practicing ear sensitivity using this game? table 7. question weight weight information likert scale 3 yes 0-33 2 maybe 34-66 1 no 67-100 table 8. beta test results question score respondent scale value results 1 83 93 (21 yes, 10 enough, 0 no) 89.25% yes 2 85 93 (23 yes, 8 enough, 0 no) 91.40% yes 3 75 93 (15 yes, 14 enough, 2 no) 80.64% yes 4 82 93 (22 yes, 7 enough, 2 no) 90.11% yes 5 80 93 (19 yes, 11 enough, 1 no) 87.91% yes from the test results obtained 89.25% of users get new knowledge related to tone interval theory, 91.40% of users say the tone played can be heard clearly, 80.64% of users feel that the training provided is in accordance with their level of expertise, 90.11% of users can understand the given tone interval theory, 87.91% of users feel comfortable practicing ear sensitivity using this educational game. https://doi.org/10.34010/injiiscom.v4i1.9588 99 | international journal of informatics information system and computer engineering 4(1) (2023) 89-100 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 4. conclusion based on the results of research, analysis, system design, and implementation and testing, it can be concluded that the game that was built can be an alternative tool that helps users or novice musicians to carry out the process of training ear sensitivity to tone intervals without involving the teacher in the learning process. applications that are built can also provide a little insight into music theory related to tone intervals. this ear sensitivity training educational game that has been built still has shortcomings, therefore the following are some acceptable suggestions for the development of this game, as follows: 1. learning and training materials can be expanded by adding other scales such as minor scales. 2. votes as a training question can be reproduced to increase difficulty and variety. acknowledgments we acknowledged bangdos, universitas pendidikan indonesia. references afrianto, i., & furqon, r. m. (2018). the herbalist game edukasi pengobatan herbal berbasis android. jurnal sistem informasi bisnis, 8(2), 141. aldwell, e., schachter, c., & cadwallader, a. (2018). harmony and voice leading. cengage learning. ardiningsih, d. (2019). pengembangan game kuis interaktif sebagai instrumen evaluasi formatif pada mata kuliah teori musik. jurnal inovasi teknologi pendidikan, 6(1), 92-103. arsenault, d. (2009). video game genre, evolution and innovation. eludamos: journal for computer game culture, 3(2), 149-176. haditama, i., slamet, c., & fauzy, d. (2016). implementasi algoritma fisher-yates dan fuzzy tsukamoto dalam game kuis tebak nada sunda berbasis android. jurnal online informatika, 1(1), 51-58. jubaedi, a. d., & putra, r. e. (2018, november). pembelajaran alat musik kendang pada kesenian beluk berbasis game. in prosiding seminar nasional rekayasa teknologi informasi| snartisi (vol. 1). laksana, d. j., budiman, a., & apriandari, w. (2021). game edukasi pengenalan alat musik tradisional menggunakan metode mdlc berbasis android. jutisi: jurnal ilmiah teknik informatika dan sistem informasi, 10(1), 45-56. levine, m. (2011). the jazz theory book. " o'reilly media, inc.". limin, s. (2022). strategi pembelajaran pada mata kuliah teori musik dengan menggunakan aplikasi kahoot. psalmoz: a journal of creative and study of church music, 3(1), 10-19. https://doi.org/10.34010/injiiscom.v4i1.9588 michael and hanhan. development of an educational training game for ear ,...| 100 doi: https://doi.org/10.34010/injiiscom.v4i1.9588 p-issn 2810-0670 e-issn 2775-5584 mursid, r. (2018). pengembangan media pembelajaran interaktif pada mata pelajaran bahasa inggris. jurnal teknologi lnformasi dan komunikasi dalam pendidika, 5(2), 210-221. pavlik, p. i., hua, h., williams, j., & bidelman, g. (2013, july). modeling and optimizing forgetting and spacing effects during musical interval training. in educational data mining 2013. pradana, a. g. (2019, october). rancang bangun game edukasi “amudra” alat musik daerah berbasis android. in prosiding seminar nasional teknologi informasi dan komunikasi (senatik) (vol. 2, no. 1, pp. 49-53). rizqyawan, m. i., & hermawan, g. (2015, october). adventure game as learning media for introducing music interval and ear training to kids. in 2015 international conference on automation, cognitive science, optics, micro electromechanical system, and information technology (icacomit) (pp. 172-175). ieee. santoso, a. m. h. (2022). covid-19: varian dan mutasi. jurnal medika hutama, 3(02 januari), 1980-1986. willis, g. (1998). ultimate ear training for guitar & bass. h. leonard. yusuf, m. (2019). peningkatan kemampuan guru dalam menentukan kriteria ketuntasan minimal (kkm) melalui workshop di uptd sdn banda soleh 1 kecamatan kokop kabupaten bangkalan tahun 2019. re-jiem (research journal of islamic education management), 2(1), 131-144. https://doi.org/10.34010/injiiscom.v4i1.9588 53 | international journal of informatics information system and computer engineering 2(2) (2021) 53-65 risks of chronic kidney disease prediction using various data mining algorithms akalya devi c*, fatima abdul jabbar**, kavi varshini s***, kriti s rithanya****, miruthubashini m*****, naveena k s****** *assistant professor, 2ug scholar, department of information technology, psg college of technology, coimbatore, india. *corresponding email: 1akalya.jk@gmail.com a b s t r a c t s a r t i c l e i n f o twenty million people have chronic kidney disease where patients experience a gradual deterioration of kidney function, the result of which is kidney failure. early detection of chronic renal disease can help to slow its progression, avert complications, and reduce the risk of cardiovascular complications. data mining has been broadly used in order to support medical professionals and physicians in the prediction and examination. here, in this paper, multiple data mining algorithms are used to solve a problem in the field of medical diagnosis and examine how effective they were at predicting the consequences. the study's focus was on the diagnosis of chronic renal disease. this dataset used for this study consists 400 instances & 25 attributes. preprocessing of the large amount of raw data is carried out to impute any missing data and determine which of the variables should be taken into account in the prediction models. the accuracy of the prediction is used to compare and contrast the various predictive analytic models. article history: received 18 dec 2021 revised 20 dec 2021 accepted 25 dec 2021 available online 26 dec 2021 aug 2018 __________________ keywords: chronic kidney disease, knearest neighbor classification, predictive analytics, decision tree, data mining, support vector machine, random forest. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(2) (2021) 53-65 mailto:akalya.jk@gmail.com akalya et al. risks of chronic kidney disease prediction using various data mining...| 54 1. introduction chronic kidney disease (ckd) or chronic kidney failure is the increasing impairment of the kidney's ability to function normally. chronic kidney disease is induced primarily due to high blood pressure, diabetes, hypertension, and several other factors in particular smoking, obesity, heart disease, heredity, consumption of alcohol, usage of drugs, age, race, ethnicity, etc. in india and other developing nations, chronic diseases remain a leading cause of mortality. the number of casualties in india owing to chronic disease was anticipated to be 5.21 million in 2008 and is likely to increase from over 7.63 million by 2020. there are five distinct stages of disease development in which each stage increases in severity while as it advances between stage 1 and stage 5. stage 1 is when a person's kidney function falls below normal. as the affected individuals it goes ahead into step 2, they may experience a mild to moderate loss of renal functions. the worsening condition escalates in level 3, where there is a moderate to average deterioration in the nephrological operation followed by acute damage in the functioning of the excretory system in stage 4. stage 5 is the absolute collapsing of the urinary organs. (almustafa, 2021). the massive increase in the amount of medical data available to predict the disease has raised the question of being effectively classified, managed, and transferred. to extract useful insight and knowledge from this raw data, effective ways are required. data mining techniques are a dependable and pragmatic way of accomplishing this. data mining is the process for processing massive amounts of data and extracting knowledge from all of this. in addition to the medical sector, the data are sequentially organized and are exploited in multiple number of real-time applications such as social networking sites, online websites, and so on. data mining is categorized in many other domains including graphic data extraction, web data mining, textual data mining, image data extraction domain. these data mining sectors facilitate in decision-making and the extraction of useful information from the dataset undergoing investigation. prediction of the risk of chronic kidney disease is based on several health parameters including random blood glucose level, blood pressure, serum creatinine level, and others. supervised classification algorithms which are used to predict the risk of chronic kidney disease are decision tree classification, support vector machine classification (svm), random forest classification, and k-nearest neighbor classification (knn) (aqlan, et al., 2017). from experimenting, random forest classification and knn were shown to be the best classifiers for classification. random forest and knn classifications have maximum reliability than decision tree and svm classifiers. 2. literatur review in this research paper, recent data mining procedures were used to classify and forecast chronic kidney disease which considers various influencing factors such as blood pressure, red blood cells count, haemoglobin, etc. the techniques used in this paper provide more accuracy than the techniques used in other existing works. 55 | international journal of informatics information system and computer engineering 2(2) (2021) 53-65 kaur g et al. applied two data mining classifiers to predict chronic kidney disease: knn and svm, which gave the exactitude and error percentage (arasu, d., & thirumalaiselvi, r. 2017). bhatla n et al. has analysed most of the dangerous diseases among which breast cancer, heart disease, and diabetes are the predominant ones (bhatla, n., & jyoti, k. 2012). on investigating 168 articles the techniques for implementing the diagnosis of various diseases have been performed. all techniques, data mining approaches, and evaluation methodologies are carefully investigated and properly considered. kunwar v et al. using categorization approaches such as naive bayes and artificial neural networks (ann), authors hypothesize chronic renal disease. according to the rapidminer tool's trial results, naive bayes generates further accurate outcomes than ann. (gharibdousti, m. s et al., 2017). decision tree, linear regression, super vector machine, naive bayesian, and artificial neural networks (anns) were one of the classification strategies utilized (ilyas, et al., 2021). the correlation matrix was used to investigate the features' correlation. as a result, they observed the influence of properties on classification findings. padmanaban k. a et al. on the incurable renal disease dataset, researchers implemented data extraction algorithms such as naive bayes and the decision tree algorithm (padmanaban, k. a., & parthiban, g. (2016). on comparing and contrasting several categorization algorithms, they recommended decision tree classification to reach substantial results with suitable accuracy by estimating its performance to its specificity and sensitivity (kunwar, et al., 2021). sharma s et al. evaluated 12 data mining clustering techniques by implementing them to the ckd dataset (sharma, et al., 2016). to determine efficiency, the findings of the prediction were contrasted with the factual medical outcomes. a few of the metrics used to evaluate performance comprise predictive accuracy, precision, sensitivity, and specificity (kunwar, et al., 2016). with an accuracy at about 98.6%, a sensitivity of 0.9720, a precision of 1, and a specificity of 1, the decision tree showed the best performance. arasu d et al. employed significant data extraction methods in particular clustering, classification, association analysis, and regression to predict renal diseases (milley, a. 2000). these techniques had insignificant shortcomings in the picturality of preprocessing or at any other stages. various data mining techniques are evaluated and the major problems are briefly explained. vijayarani s et al has focused on using a novel machine learning classification strategy to predict chronic renal illness employing svm on a data sample of 400 observations and 24 attributes (vijayarani, et al., 2015). 3. proposed work 1. due to ckd millions of individuals pass on each year since they don't experience legitimate treatment. ckd risk factors fall under four main categories: susceptibility components which lead to a rise in renal damage susceptibility, akalya et al. risks of chronic kidney disease prediction using various data mining...| 56 2. the terminology "initiation factors" refers to the elements that play a key role in renal damage. 3. progression factors leads to more regrettable reality of kidney harm and fast decay functionalities once the harm gets begun. 4. kidney failure occurs as a result of end-stage conditions, culminating in morbidity and mortality. kidney illnesses are anticipated and compared utilizing svm and ann algorithm stationed on the exactness and performance time. svm, knn, and some other algorithms have been used to assess the performance of the ckd dataset from the uci repository and the raw data which have been taken was cleaned and processed by various steps which have been explained in the figure 1. four different classifiers have been analyzed majorly established on the succeeding approaches: decision-tree, support vector machine (svm), random forest, k-nearest neighbor (knn) in section 3. these technics were picked for the examination and review for the reason that of their ubiquity within the later important writing. a concise portrayal about the chosen strategies has been given underneath. 3.1. data mining algorithms & technuques an algorithmic data mining program can be a well-specified plan of action that takes data as in and out. it includes designs in the shape of models. it comprises a small number of algorithms and strategies namely classification, grouping, prediction, association rules, neuronal networks, etc., to perform knowledge revelation from data banks. table 1 shows the evaluation plans employed here. table 1. classification of ckd & evaluation plans s. no phases of ckd gfr (glomerular filtration rate) evaluation strategies 1. nephrological damage with common gfr 90 or beyond treating the coexisting conditions, reduction of hazard variables for cardiac and vascular illness 2. renal impairment with moderate reduction 60-89 approximation of ailment advancement 3. reasonable reduction 30-59 assessment and medication of sickness intricacies 57 | international journal of informatics information system and computer engineering 2(2) (2021) 53-65 4. rigorous diminution 15-29 formulation of excretory organs switching remedy (dialysis, granting) 5. renal failure less than 15 nephrological organs grafting therapy the prediction analytics conducted is based upon the typically picked data columns of data, which comprises of the age, blood pressure, number of red blood cells, and appetite fields. these above mentioned four entries incorporate the numeric data in the case of blood pressure and age, while categorical data for the number rbgs and appetite. the nominal data has indeed been converted into numeric types so as to –make classification techniques suitable to string-based categorical attributes, which cannot be handled using statistical models. the proposed framework for the study is illustrated in figure 1. figure 1. proposed framework for ckd analysis and prediction akalya et al. risks of chronic kidney disease prediction using various data mining...| 58 here are the basic steps which were performed initially; 1. acquire the data from the local disk. 2. with the help of the column identifier ids, manually choose the columns. 3. to make all the nominal values numerical, the conversion is made. 4. after the categorical transformation, make the last data matrix. 5. inside the last data matrix, search for the missing values. 6. compute the average of every column that constitutes the variable. 7. load in the missing values with the appropriate average value from the mean values. 8. to make a non-uniform feature matrix, shuffle the data matrix. 9. divide the training and testing data matrices. 10. make the observation vectors ready for training and testing. 3.2. classification the best and most common data mining approach is classification. where entities are classified into different categories called classes and assigned to them. each and every thing needs to be distributed precisely to one class and not more than one and never to no classes at all. decision tree, svm, knn, and random forest were the classification algorithms included in this model. 3.2.1. decision tree classification this method is especially beneficial for deciphering classification problems in which a tree is formed to depict the categorization process. the tree is linked to every tuple in the database to yield classification as long as it is established. classification tree analysis and regression tree analysis are the two forms of decision trees used in data mining, and they have been employed for a spectrum of potential results such as belonging to a specific statistical class or an actual number. 1. fitting decision tree to the training set. 2. predicting the test result. 3. calculating the accuracy. 4. displaying the confusion matrix. 3.2.2 svm classification svm is a set of rules for supervised machine learning that can be used to resolve classification and regression problems. it uses a strategy called the kernel trick to convert your data and after that based on these changes it finds an ideal boundary between the possible outputs (sinha, p., & sinha, p. 2015). the following steps are the ones performed; 1. support vector machine (a classification technique) is applied on the available data for the purpose of predictive analysis. 2. using the training data matrix and the training observation vector the classifier is trained. 3. the testing data matrix with unseen data is utilized to examine the classifier 59 | international journal of informatics information system and computer engineering 2(2) (2021) 53-65 4. the predictions (observations predicted by svm classifier) are returned as output. the entire performance is computed by comparing and contrasting the outcomes of support vector machine classifier and the actual perceptions. 3.2.3 random – forest classifier random forest is an analyzer that equips the average of a number of decision trees on discrete subsections of a given set of data to advance the dataset's predicted accuracy. the following steps are the ones performed (sinha, p., & sinha, p. 2015). 1. fitting random forest classification to the training set. 2. predicting the test result. 3. calculating the accuracy. 4. displaying the confusion matrix. 3.2.4 knn classifier it's a type of distance-based technique that's typically used while the values of each and every attribute is uninterrupted and continual, but it can also be used with nominal features (subasi, et al., 2017). to compute the categorization of an unknown sample data based on the classification of the closest instance or instances. more occurrences inside the preparation set use the same way to group the k-nearest neighbors (also known as k-nearest neighbor), (vijayarani, et al., 2015). the steps that were taken were as follows: 1. k-nearest neighbor (one of the classification technics) is employed over the given data for the purpose of predictive analytics. 2. before initiating the entire process, the value of k should be initialized which will be symbolizing the number of neighbors that has to be considered. 3. the k-nearest neighbor classifier needs to be trained with the specified k value over the training data matrix and the training observation vector. 4. with the help of the test data matrix, which contains the unseen data the classifier is tested and evaluated for the required metrics. 5. the forecasts (observations predicted by the knn classifier) made by the knn analyzer should be returned. 6. the entire accuracy and performance of the knn classifier is estimated by comparing the predictions made by knn and the actual observations. 5. result and disscusion the chronic kidney disease (ckd) dataset was acquired based on the uci machine learning repository and is employed in this study for prediction and validation. both numerical and nominal attributes were included in ckd dataset. there are 25 attributes and 400 instances. this dataset also contains missing values. there are 24 attributes and one class attribute (i.e.) ckd, not-ckd. table 2 gives the attribute description of the dataset. akalya et al. risks of chronic kidney disease prediction using various data mining...| 60 table 2. ckd dataset attributes description s. no attribute name expansion s. no attribute name expansion 1 age age of the patient 13 pot potassium 2 bp blood pressure 14 hemo hemoglobin 3 sg specific gravity 15 pcv packed cell volume 4 al albumin 16 wc white blood cell count 5 su sugar 17 rc red blood cell count 6 rbc red blood cells 18 htn hypertension 7 pc pus cell 19 dm diabetes mellitus 8 pcc pus cell clumps 20 cad coronary artery disease 9 ba bacteria 21 appet appetite 10 bgr blood glucose random 22 pe pedal edema 11 bu blood urea 23 ane anemia 12 sc serum creatinine 24 sod sodium data cleaning and data preprocessing is the most critical point in the data mining procedure as it influences the rate of success drastically. the categorical attributes were displaced with 0s and 1s corresponding to their values. the missing values were replaced with the mean of that particular attribute. as there was a wide range of age, the age attribute was grouped in batches (sharma, et al., 2016). the ckd dataset includes features that vary in the degree of magnitude, range, and units. in order to interpret all the features on the same scale, feature scaling (data normalization) was carried out. the ckd dataset was parted into 70% for the purpose training and 30% for the purpose of testing data. four different data mining procedures encompassing decision tree classification, support vector machine classification, random forest classification, knn classification 61 | international journal of informatics information system and computer engineering 2(2) (2021) 53-65 were applied to the training and testing data and the performance measurement using different metrics like precision, f1score, recall, accuracy, specificity, and sensitivity were observed (vijayarani, et al., 2015). table 3 presents the different performance metrics used in this paper. table 3. different overall performance analysis metrics used metrics definition equation precision the proportion of predicted accurately positive considerations to fully predicted positive observations is referred as precision. recall (sensitivity) estimates the percentage of number of yes’s that are effectively-recognized correctly. f1-score precision and recall are weighted averages which determine the f1 score. accuracy measures the model's ability to accurately estimate class label of latest or previously unknown information. specificity here, ratio of negatives (or no's) that have been correctly recognized as such is measured. the performance metrics of the various proposed algorithms were derived using the equations listed in table 3. table 4 depicts the results obtained for every algorithm. table 4. performance measures of the proposed algorithms model precision recall f1-score specificity accuracy not-ckd ckd not-ckd ckd not-ckd ckd decision tree 0.91 1.00 1.00 0.95 0.95 0.97 1.00 0.967 svm 0.93 1.00 1.00 0.96 0.97 0.98 1.00 0.975 akalya et al. risks of chronic kidney disease prediction using various data mining...| 62 model precision recall f1-score specificity accuracy not-ckd ckd not-ckd ckd not-ckd ckd random forest 0.95 0.99 0.98 0.97 0.96 0.98 0.987 0.975 knn 0.95 1.00 1.00 0.97 0.98 0.99 1.00 0.983 the train score is the measurement that states us in what way the model suits the training data. similarly, the test score shows how the model reacts to the unknown data. the area under the curve (auc) score portrays the model’s overall performance at differentiating between the positive and negative classes. figure 2 shows the comparison of the training score, test score and mean auc score. figure 2. depiction of train, test, and mean auc scores of the proposed algorithms the difference in magnitude between both the observation's prediction and its true value is termed as the mean absolute error. for the proposed algorithms, figure 3 illustrates the mean absolute error. 63 | international journal of informatics information system and computer engineering 2(2) (2021) 53-65 figure 3. illustration of the proposed algorithms' mean absolute error the receiver operating characteristic curve (roc curve) reflects the classification model's overall performance among all class thresholds [10]. the roc curve for the algorithms employed in this study is illustrated in figure 4. figure 4. plot of roc curve for the proposed algorithms 5. conclusion the objective of this article is to analyze the variety of data mining techniques and algorithms utilized to predict chronic renal disease. ckd has been predicted and diagnosed using data akalya et al. risks of chronic kidney disease prediction using various data mining...| 64 mining classifiers: decision tree, svm, random forest, and knn. it was found that knn results in the best accuracy. the performance of the knn method was found to be 98.3% accurate compared to decision tree (96.7%), svm (97.5%), and random forest (97.5%). the work can further be extended keeping into consideration the other parameters like food intake, living conditions like sanitation, availability of clean water, working environment, environmental factors like pollution, etc. for the detection of kidney disease. further experimentation can be conducted using other classifiers like ann or by using ensemble techniques. references almustafa, k. m. (2021). prediction of chronic kidney disease using different classification algorithms. informatics in medicine unlocked, 100631.dobrucka, r. (2018). synthesis of mgo nanoparticles using artemisia abrotanum herba extract and their antioxidant and photocatalytic properties. iranian journal of science and technology, transactions a: science, 42(2), pp. 547-555. aqlan, f., markle, r., & shamsan, a. (2017). data mining for chronic kidney disease prediction. in iie annual conference. proceedings (pp. 1789-1794). institute of industrial and systems engineers (iise). arasu, d., & thirumalaiselvi, r. (2017). review of chronic kidney disease based on data mining techniques. international journal of applied engineering research, 12(23), 13498-13505. bhatla, n., & jyoti, k. (2012). an analysis of heart disease prediction using different data mining techniques. international journal of engineering, 1(8), 1-4 gharibdousti, m. s., azimi, k., hathikal, s., & won, d. h. (2017). prediction of chronic kidney disease using data mining techniques. in iie annual conference. proceedings (pp. 2135-2140). institute of industrial and systems engineers (iise). ilyas, h., ali, s., ponum, m., hasan, o., mahmood, m. t., iftikhar, m., & malik, m. h. (2021). chronic kidney disease diagnosis using decision tree algorithms. bmc nephrology, 22(1), 1-11. ilyas, h., ali, s., ponum, m., hasan, o., mahmood, m. t., iftikhar, m., & malik, m. h. (2021). chronic kidney disease diagnosis using decision tree algorithms. bmc nephrology, 22(1), 1-11. kunwar, v., chandel, k., sabitha, a. s., & bansal, a. (2016, january). chronic kidney disease analysis using data mining classification techniques. in 2016 6th international conference-cloud system and big data engineering (confluence) (pp. 300-305). ieee. 65 | international journal of informatics information system and computer engineering 2(2) (2021) 53-65 milley, a. (2000). healthcare and data mining. health management technology, 21(8), 44-45. padmanaban, k. a., & parthiban, g. (2016). applying machine learning techniques for predicting the risk of chronic kidney disease. indian journal of science and technology, 9(29), 1-6. sharma, s., sharma, v., & sharma, a. (2016). performance based evaluation of various machine learning classification techniques for chronic kidney disease diagnosis. arxiv preprint arxiv:1606.09581. sinha, p., & sinha, p. (2015). comparative study of chronic kidney disease prediction using knn and svm. international journal of engineering research and technology, 4(12), 608-12. subasi, a., alickovic, e., & kevric, j. (2017). diagnosis of chronic kidney disease by using random forest. in cmbebih 2017 (pp. 589-594). springer, singapore. rubini, l. j., & eswaran, p. (2015). uci machine learning repository: chronic_kidney_disease data set. vijayarani, s., dhayanand, s., & phil, m. (2015). kidney disease prediction using svm and ann algorithms. international journal of computing and business research (ijcbr), 6(2), 1-12. 75 | international journal of informatics information system and computer engineering 4(1) (2023) 75-88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 a computational bibliometric analysis of e-groceries analysis using vosviewer rudhi lesmana*, m ihsan rifaldi departemen manajemen, universitas komputer indonesia, indonesia *corresponding email: rudhi.21221230@mahasiswa.unikom.ac.id a b s t r a c t s a r t i c l e i n f o the purpose of the research is to combine mapping analysis with vosviewer as well as publish or perish software to do a computerized bibliometric analysis of the topic "e-groceries analysis." the method used descriptive-quantitative approach in conjunction with bibliometric analysis in which the data were retrieved from google scholar. based on the results, e-groceries analysis research decreases every year, proven by the fact that 2018 have 25 articles and increased to 32 articles in 2019, 49 articles in 2020, and 98 articles in 2021. based on further findings of this research, it can be concluded that there are several understudied sectors in e groceries analysis that may be examined further to increase the efficacy of e-groceries analysis. this research is also anticipated to serve as a reference for future research in defining and assessing the research subject, as well as a reference for field to be studied in e-groceries analysis. article history: submitted/received 01 oct 202 2 first revised 15 jan 2023 accepted 03 mar 2023 first available online 14 apr 2023 publication date 01 jun 2023 aug 2018 __________________ keywords: bibliometrics, e-groceries analysis, data analysis, vosviewer international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(1) (2023) 75-88 https://doi.org/10.34010/injiiscom.v4i1.9586 mailto:rudhi.21221230@mahasiswa.unikom.ac.id rudhi & m. ihsan. a computational bibliometric analysis of e-groceries...| 76 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 1. introduction the customary sequences that were once employed when completing daily tasks, including shopping, have been disturbed and re-combined, both in time and place, as a result of the internet today (couclelis, h. 2000). online shopping does certainly allow customers to buy goods or services from a seller over the internet, fundamentally changing the procedures involved in information gathering, comparison, and use, as well as purchase and delivery. people who use e-commerce can purchase products using their mobile devices while, for instance, traveling to work or waiting at the train station without having to adhere to the store's precise opening and closing hours. consumer behavior is significantly altered by the evolution of shopping, and this behavior is closely related to transportation (suel, e., & polak, j. w. 2017). with grocery shopping being the most popular and regular form of retail therapy, it has a particularly negative impact on the environment and urban freight transportation. however, depending on customer behaviour and last mile delivery strategies, switching from in-store to online purchasing can have both good and bad effects on transportation. in greater detail, it is evident that when customers order groceries online and want home delivery, the burden of the freight travels is transferred from the customer to the store. instead, the final effect on urban freight transportation is unpredictable because it relies on the kind of product, how often people shop, why they purchase, whether trips are chained together, and how quickly efficiency must be achieved (mokhtarian, p. l. 2004). therefore, this study aims to conduct a bibliometric analysis on the topic of purchasing decisions in using the egroceries service. this method uses a mixed method with a literature review, publish or perish to collect data and vosviewer to visualize the relationship between terms as well as other things such as research trends throughout the year. it is hoped that this research will contribute to finding the fields proposed in the topic of e-groceries analysis. egroceries analysis is a business model that applies information technology to establish communication relationships and conduct transactions with customers regarding products, services and distribution systems through internet media (muhammad, n. s., et al 2016). previous study regarding e-groceries analysis have been conducted. ayudhia et al. conducted a study regarding egroceries analysis of business model. pico and barcelo also conducted a study regarding e-groceries study, which focuses on organic matter and microplastics. according to pico and barcelo, py-gc-ms is a valuable technique for e-groceries analysis specially to cover crucial e-groceries aspects (pico, y., & barcelo, d. 2020). besides, plenty of bibliometric analysis research on various fields, such as computer science (al husaeni, d. f., & nandiyanto, a. b. d. 2022), educational research (al husaeni, d. f., et al 2023), high school (al husaeni, d. n., & nandiyanto, a. b. d. 2023), technoeconomic education (ragadhita, r., & nandiyanto, a. b. d. 2022), materials research (nandiyanto, a. b. d., et al 2020), vocational school (al husaeni, d. https://doi.org/10.34010/injiiscom.v4i1.9586 77 | international journal of informatics information system and computer engineering 4(1) (2023) 75-88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 n., & nandiyanto, a. b. d. 2023), digital learning (al husaeni, d. f., & nandiyanto, a. b. d. 2022), scientific publications (mulyawati, i. b., & ramadhan, d. f. 2021), bioenergy management (soegoto, h., et al., 2022), chemical engineering (nandiyanto, a. b. et al 2021), special needs education (al husaeni, d. f., et al., 2023), and covid-19 (hamidah, i., sriyono, s., & hudha, m. n. 2020). however, there have not been a bibliometric analysis regarding egroceries analysis. therefore, this research aims to conduct a bibliometric analysis on the topic of egroceries analysis. the method used mixed method with literature review, publish or perish 8 to gather the data and vosviewer to visualize the connection between terms as well as other things such as research trend along the year. it is hoped that this research would contribute to discover the understudied fields in the topic of e-groceries analysis. 2. method descriptive-quantitative approaches were applied in this study. in addition, literature review were conducted to gain insights based on previous researches on bibliometric analysis as well as the topic of e-groceries analysis. we collected the articles from journals indexed by google scholar, due to its accessibility. publish or perish was chosen to gather the bibliometric data from google scholar (al husaeni, d. f., & nandiyanto, a. b. d. 2022). then, the bibliometric data were saved in *.ris, and *.csv format to be used in vosviewer software and to be converted into *.xlsx to be analyzed further. the software version that is used in this research is publish or perish 8 and vosviewer 1.6.17. in this research, we sifted through facts and used relevant facts to make arguments under the topic e-groceries analysis. we retrieve the data from google scholar by entering the keyword "e-groceries analysis" for to the title, keyword, and abstract requirements in the publish or perish software. we obtained 993 articles on e-groceries analysis research published between 2017 and 2021. the collected articles are then saved in *.ris format to be visualized in vosviewer software in the form of visualization map, and to analyze the research trend in the form of bibliometric maps. before creating the map, irrelevant terms were filtered in the visualization map (allan, r. n., et al., 1984). the visualization map is classified into three types: network visualization, overlay visualization, and density visualization. 3. results and discussion 3.1. research developments in the field of e-groceries analysis research on climate development in the field of e-groceries analysis, describes the development of research in the field of e-groceries analysis from 2018 to 2021 in fig. 1. figure 1 shows that the research on egroceries analysis decreases every year. this can be proven by the fact that there are 25 articles in 2018, 32 articles in 2019, 49 articles in 2020, and lastly 98 articles in 2021. based on the search results in the publish or perish, there are 263 articles that match the research topic. 16 articles with the most citations from 16 different publishers were shown in table 1. https://doi.org/10.34010/injiiscom.v4i1.9586 rudhi & m. ihsan. a computational bibliometric analysis of e-groceries...| 78 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 fig. 1. level of research development on e-groceries analysis table 1. article data in the field of e-groceries analysis no authors title publisher year cites refs 1. oa hjelkre m., et al. e-groceries: sustainable last mile distribution in city planning wiley online library 2021 255 (bjørgen, a., et al., 2021) 2. c fikar a decision support system to investigate food losses in egrocery deliveries westminst erresearch. westminst er.ac 2018 63 (fikar, c. 2018) 3. by ekren., et al. lateral inventory sharebased models for iot-enabled e-commerce sustainable food supply networks university of jaffna 2021 57 (ekren, b. y., et al., 2021) https://doi.org/10.34010/injiiscom.v4i1.9586 79 | international journal of informatics information system and computer engineering 4(1) (2023) 75-88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 table 1 (continue). article data in the field of e-groceries analysis no authors title publisher year cites refs 4. c thommi s logistieke uitdagingen in e-groceries uis.brage.u nit.no 2021 52 (thommis, c. 2021) 5. m fernand ez vazque znoguer ol. modeling and optimization of the supply chain in e-groceries uir.unisa.a c.za 2021 46 (fernande z v, n, m. 2021) 6. m mees. e-groceries: the effects of simulated sensory information and freshness guarantee information on consumer uncertainty. uijrt.com 2019 44 (mees, m. 2019) 7. c berggre n., & s wikströ m barriers online: exploring consumers' resistance to egroceries ubiblioru m.ubi.pt 2018 43 (berggren, c., & wikström, s. 2018) 8. ai pujol digital nudging to enhance sustainable purchasing behaviours in egroceries turcomat.o rg 2020 42 (pujol, a. i. 2020) 9. mfv noguer ol modeling and optimization of the supply chain in e-groceries thesis.cust. edu.pk 2021 40 (noguerol, m. f. v. 2021) https://doi.org/10.34010/injiiscom.v4i1.9586 rudhi & m. ihsan. a computational bibliometric analysis of e-groceries...| 80 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 table 1 (continue). article data in the field of e-groceries analysis no authors title publisher year cites refs 10. j meijboo m waste reduction in e-groceries fulfilment center: a case study at picnic theseus.fi 2019 38 (meijboom , j. 2019) 11. p gunawa rdana., & pin fernand o does customer trust mediate the impact of eservice quality dimensions? lessons during covid-19 pandemic tesi.luiss.it 2021 30 (gunawar dana, p. k. a. t. d. r., & fernando, i. 2021) 12. p gunawa rdana., & pin fernand o does customer trust impact on e-service quality dimensions during covid-19 pandemic? iopscience. iop.org 2021 29 (gunawar dana, p. k. a. t. d. r., & fernando, i. 2021) 13. p gunawa rdana., & pin fernand o assessing the mediation role of the customer trust on eservice quality: lessons during covid-19 pandemic cambridge university press 2021 28 (gunawar dana, p. k. a. t. d. r., & fernando, i. 2021) 14. y kusna di., & g pan developing online business strategy with millennials through partnership with university snejournal.org 2020 15 (kusnadi, y., & pan, g. 2020) https://doi.org/10.34010/injiiscom.v4i1.9586 81 | international journal of informatics information system and computer engineering 4(1) (2023) 75-88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 table 1 (continue). article data in the field of e-groceries analysis no authors title publisher year cites refs 15. vc echrler., et al. challenges and perspectives for the use of electric vehicles for last mile logistics of grocery ecommerce– findings from case studies in germany sljmuok.slj ol.info 2021 15 (ehrler, v. c., et al., 2021) 16. m waitz., et al. a decision support system for efficient lastmile distribution of fresh fruits and vegetables as part of egrocery operations search.pro quest.com 2018 13 (waitz, m., mild, a., & fikar, c. 2018) in table 1 there are 16 articles that match the criteria research. of the 16 selected articles, showing that highest quote related to e-groceries analysis research is 255, while with the lowest citation is 13. that in table 1, it shows that in 2018 and 2021, each has articles with quotes highest. in 2018-2021, the most articles quoted is 255 articles. temporary that, in 2018, a lot of articles quoted are 63 articles. year with quote the most is in 2021 as many as 255 articles. 3.2. visualization e-groceries analysis topic area using vosviewer visualization map of e-groceries analysis topic was created using vosviewer software. according to al husaeni and nandiyanto, two terms set are the minimum number of relationships when creating map using vosviewer software (peters, c. i. 1975). the generated map has 10 items (terms) with a total of 3 clusters, 18 links, and total link strength of 166 (see fig. 2). cluster 1 is indicated by red; cluster 2 is shown in green; cluster 3 is shown in dark blue. figure 2 is the network visualization map generated by vosviewer based on the terms present in collected data. the collected articles have a total of 10 terms (in the form of items) and were categorized into 3 clusters. in addition, it has the total link strength of 166 and total links of 18. the item categorization is determined based on the connection https://doi.org/10.34010/injiiscom.v4i1.9586 rudhi & m. ihsan. a computational bibliometric analysis of e-groceries...| 82 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 strength of the terms with each other, further detail of each cluster is shown in figs. 3 7. items on each cluster are as follows: (i) cluster 1 (4 items) customer, e-grocery, home delivery, supply chain (ii) cluster 2 (3 items) feature, main content skip, skip (iii) cluster 3 (3 items) covid, pandemic, role fig. 2. network visualization map of e-groceries analysis fig. 3. cluster 1 visualization e-groceries analysis network. https://doi.org/10.34010/injiiscom.v4i1.9586 83 | international journal of informatics information system and computer engineering 4(1) (2023) 75-88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 the main node in cluster 1 is the term ‘egroceries’, this node linked to several other nodes in cluster 1 namely, ‘supply chain’, ‘costumer’, and ‘home delivery’. in addition, it also linked to the nodes in the other cluster, such as • 'supply chain', 'costumer', and 'home delivery' in cluster 1 • 'feature', 'main content skip', and 'skip' in cluster 2 • 'covid', 'pandemic', and 'role' in cluster 3 the main node in cluster 2 is the term, egroceries feature, this node linked to several other nodes in cluster 2 namely, ‘skip’, and ‘main content skip’. in addition, it also linked to the nodes in the other cluster, such as • 'supply chain', 'costumer', and 'home delivery' in cluster 1 • 'feature', 'main content skip', and 'skip' in cluster 2 • ' covid', 'pandemic', and 'role' in cluster 3 fig. 4. cluster 2 visualization e-groceries analysis network. https://doi.org/10.34010/injiiscom.v4i1.9586 rudhi & m. ihsan. a computational bibliometric analysis of e-groceries...| 84 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 fig 5. cluster 3 visualization e-groceries analysis network. 3.3. overlay visualization map of egroceries analysis overlay visualization map visualize the research trend of keywords in each year. different coloration indicates the year in which terms are commonly used. darker color indicates that the keyword is commonly appear on older years while bright color indicates that the keyword commonly appears on recent year. in fig. 6, the majority of keywords seems to be popular on older years. however, there are recently emerging keywords in the collected data such as ‘covid’, ‘egrocery’, ‘customer’, ‘feature’. these keywords can be linked to recent situations such as the covid-19 pandemic and the effort to minimize carbon footprint and green energy development in the name of saving the environment. 3.4. density visualization of e-groceries analysis density visualization aims to show the frequency of occurrence of terms in the collected data. color intensity and size is the primary indicator of density, so an item that have a large and bright coloration means that the keyword appears frequently in the collected data and vice versa. the density visualization is shown in fig. 7. visualization density about climate egroceries analysis research is in the picture above, which means that on the map density showing results analysis use all article regarding e-groceries analysis in 2018-2022. in fig. 7, it is depicted that there are some color terms that is there is color yellow with a fairly large diameter. these terms called evidence, e-grocery, covud, customer, and feature. https://doi.org/10.34010/injiiscom.v4i1.9586 85 | international journal of informatics information system and computer engineering 4(1) (2023) 75-88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 fig. 6. overlay e-groceries analysis visualization fig. 7. density visualization map of e-groceries analysis 4. conclusion the conclusion in this study is that there are many topics that are poorly explored in the field of e-groceries analysis for example, cluster 1 is "e-groceriey", cluster 2 "feature", cluster 3 "covid". it is hoped that this research will contribute to finding the field studied in the topic of egroceries analysis references al husaeni, d. f., & nandiyanto, a. b. d. (2022). bibliometric using vosviewer with publish or perish (using google scholar data): from step -by-step processing for users to the practical examples in the analysis of digital learning articles in pre https://doi.org/10.34010/injiiscom.v4i1.9586 rudhi & m. ihsan. a computational bibliometric analysis of e-groceries...| 86 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 and post covid-19 pandemic. asean journal of science and engineering, 2(1), 1946. al husaeni, d. f., & nandiyanto, a. b. d. (2022). bibliometric using vosviewer with publish or perish (using google scholar data): from step -by-step processing for users to the practical examples in the analysis of digital learning articles in pre and post covid-19 pandemic. asean journal of science and engineering, 2(1), 1946. al husaeni, d. f., & nandiyanto, a. b. d. (2022). mapping visualization analysis of computer science research data in 2017-2021 on the google scholar database with vosviewer. international journal of informatics, information system and computer engineering (injiiscom), 3(1), 1-18. al husaeni, d. f., nandiyanto, a. b. d., & maryanti, r. (2023). bibliometric analysis of educational research in 2017 to 2021 using vosviewer: google scholar indexed research. indonesian journal of teaching in science, 3(1), 1-8. al husaeni, d. f., nandiyanto, a. b. d., & maryanti, r. (2023). bibliometric analysis of educational research in 2017 to 2021 using vosviewer: google scholar indexed research. indonesian journal of teaching in science, 3(1), 1-10. al husaeni, d. n., & nandiyanto, a. b. d. (2023). bibliometric analysis of high school keyword using vosviewer indexed by google scholar. indonesian journal of educational research and technology, 3(1), 1-12. al husaeni, d. n., & nandiyanto, a. b. d. (2023). bibliometric analysis of high school keyword using vosviewer indexed by google scholar. indonesian journal of educational research and technology, 3(1), 1-12. allan, r. n., billinton, r., & lee, s. h. (1984). bibliography of the application of probability methods in power system reliability evaluation 1977-1982. ieee power engineering review, (2), 24-25. berggren, c., & wikström, s. (2018). barriers online: exploring consumers' resistance to e-groceries. bjørgen, a., bjerkan, k. y., & hjelkrem, o. a. (2021). e-groceries: sustainable last mile distribution in city planning. research in transportation economics, 87, 100805. couclelis, h. (2000). from sustainable transportation to sustainable accessibility: can we avoid a new tragedy of the commons?. information, place, and cyberspace: issues in accessibility, 341-356. https://doi.org/10.34010/injiiscom.v4i1.9586 87 | international journal of informatics information system and computer engineering 4(1) (2023) 75-88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 ehrler, v. c., schöder, d., & seidel, s. (2021). challenges and perspectives for the use of electric vehicles for last mile logistics of grocery e -commerce–findings from case studies in germany. research in transportation economics, 87, 100757. ekren, b. y., mangla, s. k., turhanlar, e. e., kazancoglu, y., & li, g. (2021). lateral inventory share-based models for iot-enabled e-commerce sustainable food supply networks. computers & operations research, 130, 105237. fernandez vazquez-noguerol, m. (2021). modeling and optimization of the supply chain in e-groceries (doctoral dissertation, organización de empresas e márketing). fikar, c. (2018). a decision support system to investigate food losses in e -grocery deliveries. computers & industrial engineering, 117, 282-290. gunawardana, p. k. a. t. d. r., & fernando, i. (2021). does customer trust mediate the impact of e-service quality dimensions? lessons during covid-19 pandemic (preprint). gunawardana, p. k. a. t. d. r., & fernando, i. (2021). does customer trust mediate the impact of e-service quality dimensions? lessons during covid-19 pandemic. gunawardana, p. k. a. t. d. r., & fernando, p. i. n. (2021). assessing the mediation role of the customer trust on e-service quality: lessons during covid-19 pandemic. sri lanka journal of marketing, 7(3), 105. hamidah, i., sriyono, s., & hudha, m. n. (2020). a bibliometric analysis of co vid-19 research using vosviewer. indonesian journal of science and technology, 34-41. kusnadi, y., & pan, g. (2020). developing online business strategy with millennials through partnership with university. mees, m. (2019). e-groceries: the effects of simulated sensory information and freshness guarantee information on consumer uncertainty. meijboom, j. (2019). waste reduction in e-groceries fulfilment center: a case study at picnic. mokhtarian, p. l. (2004). a conceptual analysis of the transportation imp acts of b2c e-commerce. transportation, 31, 257-284. muhammad, n. s., sujak, h., & abd rahman, s. (2016). buying groceries online: the influences of electronic service quality (eservqual) and situational factors. procedia economics and finance, 37, 379-385. https://doi.org/10.34010/injiiscom.v4i1.9586 rudhi & m. ihsan. a computational bibliometric analysis of e-groceries...| 88 doi: https://doi.org/10.34010/injiiscom.v4i1.9586 p-issn 2810-0670 e-issn 2775-5584 mulyawati, i. b., & ramadhan, d. f. (2021). bibliometric and visualized analysis of scientific publications on geotechnics fields. asean journal of science and engineering education, 1(1), 37-46. nandiyanto, a. b. d., al husaeni, d. n., & al husaeni, d. f. (2021). a bibliometric analysis of chemical engineering research using vosviewer and its correlation with covid-19 pandemic condition. journal of engineering science and technology, 16(6), 4414-4422. nandiyanto, a. b. d., girsang, g. c. s., maryanti, r., ragadhita, r., anggraeni, s., fauzi, f. m., ... & al-obaidi, a. s. m. (2020). isotherm adsorption characteristics of carbon microparticles prepared from pineapple peel waste. communications in science and technology, 5(1), 31-39. noguerol, m. f. v. (2021). modeling and optimization of the supply chain in e-groceries (doctoral dissertation, universidade de vigo). peters, c. i. (1975). method of antenna tuning. department of the navy washington dc. pico, y., & barcelo, d. (2020). pyrolysis gas chromatography-mass spectrometry in environmental analysis: focus on organic matter and microplastics. trac trends in analytical chemistry, 130, 115964. pujol, a. i. (2020). digital nudging to enhance sustainable purchasing behaviours in e groceries. ragadhita, r., & nandiyanto, a. b. d. (2022). computational bibliometric analysis on publication of techno-economic education. indonesian journal of multidiciplinary research, 2(1), 213-220. soegoto, h., soegoto, e. s., luckyardi, s., & rafdhi, a. a. (2022). a bibliometric analysis of management bioenergy research using vosviewer application. indonesian journal of science and technology, 7(1), 89-104. suel, e., & polak, j. w. (2017). development of joint models for channel, store, and travel mode choice: grocery shopping in london. transportation research part a: policy and practice, 99, 147-162. thommis, c. (2021). logistieke uitdagingen in e-groceries. waitz, m., mild, a., & fikar, c. (2018). a decision support system for efficient lastmile distribution of fresh fruits and vegetables as part of e -grocery operations. https://doi.org/10.34010/injiiscom.v4i1.9586 241 | international journal of informatics information system and computer engineering 4(1) (2023) 96 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 intention to adopt cloud-based e-learning in nigerian educational institutions tom a. m 1, virgiyanti, w 2* 1school of computing, universiti utara malaysia, malaysia 2faculty of ocean engineering technology and informatics, universiti malaysia terengganu, malaysia *corresponding email: wiwied.virgiyanti@umt.edu.my a b s t r a c t s a r t i c l e i n f o institutions of higher education must utilize innovative information and communication technologies for teaching in nigeria. thus, cloud-based e-learning is essential to curtail educational challenges such as limited infrastructure, funds, and student-to-lecturer ratio. recently, there has been widespread enthusiasm regarding cloud computing for e-learning; adopting and strategically utilizing these technologies remains a significant challenge for higher education institutions. furthermore, there is a limited understanding of how cloud-based e-learning can transform nigerian educational establishments. cloud-based e-learning systems' technological components have been the subject of numerous study studies, but little is known about how they operate from an organizational perspective. accordingly, using the technologyorganization-environment theory, the goal of this study is to investigate the variables that influence the adoption of cloud-based e-learning. the findings of the research show that relative benefit and competing pressure have a big impact on whether cloud-based elearning is adopted. however, compatibility, security, and top management commitment do not appear to be significant determinants. these findings will help nigerian education institutions, the ministry of education, and practitioners to understand the critical factors for adopting this technology for improved education. article history: submitted/received 31 jul 2022 first revised 26 aug 2022 accepted 30 sept 2022 available online 29 oct 2022 publication date 01 dec 2022 aug 2018 __________________ keywords: intention to adopt, cloud-based, e-learning, higher education institution international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 241-250 https://doi.org/10.34010/injiiscom.v3i2.9563 mailto:wiwied.virgiyanti@umt.edu.my tom and virgiyanti. intention to adopt cloud-based e-learning in nigerian …| 242 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 1. introduction in recent years, the advancement of technology has brought about innovations like cloud computing, big data, ai, and blockchain (tom, virgiyanti, & rozaini, 2019). these innovations brought more opportunities and challenges to businesses. e-learning is a critical driver for emancipation from poverty (world bank [wb], 2013). research has shown that the e-learning systems of emerging nations are experiencing a similar and severe crisis. this crisis is primarily ascribed to the lack of clear education policies, infrastructure deficiencies, and inadequate investment in the education sector. the danger this scenario poses to the continent's ability to learn and advance generally is limitless. inadequate allocation of financial resources has been identified as nigeria's most significant challenge in education (asiyai, 2013; edomwonyi & osarumwense, 2017; virgiyanti & rozaini, 2019). the lack of concrete education policies, infrastructure gaps, and underinvestment in the education sector are mainly to blame for the comparable and pervasive problem in emerging nations' e-learning systems. the learning process and the continent's development as a whole are constantly threatened by this scenario. the biggest barrier to education in nigeria has been recognized to be the inadequate supply of money means. previous research has demonstrated the prevalence of issues in e-learning systems in developing countries, which can be attributed to the shortage of feasible and effective educational policies, insufficient infrastructure, and inadequate investment. the biggest obstacle to education in nigeria has been recognized as a lack of funding. (asiyai, 2013; edomwonyi & osarumwense, 2017; virgiyanti & rozaini, 2019). many universities in developing countries currently operate unreliable elearning systems. as a solution, cloud computing, with its location independence, can provide staff and students in nigerian universities with access to highly dependable and efficient systems, similar to those found in developed countries. this would ultimately enhance the competitiveness of nigerian higher education institutions (heis). to examine the usage of cloud-based e-learning from the managerial viewpoint of heis, more actual data is needed. in this research, the opinion of administrators in nigerian heis is investigated using the elements of technology, organization, and environment as well as exterior variables. the results of this research will help understand the critical elements influencing the uptake and usage of cloud-based e-learning in nigeria. 2. theoretical and conceptual model the technology, organization, and environment (toe) hypothesis, developed by tornatzky and fleischer, is used in this research (tornatzky & fleischer, 1990). the toe is an organizational theory that thoroughly explains the likelihood that a company or group will embrace innovation. the toe proposes that these three variables affect how organizations perceive the need for innovation and adapt/adopt it in order to stay competitive by including both limitations that act as roadblocks and https://doi.org/10.34010/injiiscom.v3i2.9563 243 | international journal of informatics information system and computer engineering 3(2) (2022) 241-250 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 opportunities that serve as incentives for innovation (baker, 2012; virgiyanti & rozaini, 2019). relative advantage (ra) and compatibility are technical factors when implementing cloud computing, including cloud-based e-learning platforms. (com). information and communication technology (ict) is used by universities all over the globe to create efficient and effective learning environments for both employees and pupils. as a result, when implementing innovation, implementers must weigh the advantages and disadvantages of the technology. relative advantage (ra) is the expectation of an organization's gain from technical factors. hence, institutions consider adopting innovation and its advantages over their existing systems. cloud computing has the edge over traditional server-based systems due to its flexibility, mobility, and scalability. adopting the cloud in academia opens up numerous avenues such as collaboration, discussion, availability of resources, and cost savings due to its payper-use model (tom, virgiyanti, & osman, 2019). we, therefore, posit that: 2.1. h1: relative advantage will positively influence the adoption of cloud-based e-learning in nigerian heis. due to a lack of resources, developing nations, especially those in africa, face difficulties in enhancing their educational facilities and need assistance to fix their outdated, ineffective systems. the degree to which an invention is viewed as compatible with the users' established beliefs, standards, and experiences is also referred to as compatibility (rogers, 1995). as in developed countries, cloud computing supports a variety of apps and computer languages that can be easily incorporated into nigerian e-learning systems, giving users an edge in terms of freedom and productivity. hence, we posit that: 2.2. h2: compatibility will positively influence the adoption of cloudbased e-learning in nigerian heis. regarding organizations considering implementing cloud computing, security is a top worry because it could present challenges. since data proprietorship is still a contentious topic in the context of cloud computing, the secrecy, stability, and accessibility of an organization's data are extremely important. 2.3. h3: security will positively influence the adoption of cloudbased e-learning in nigerian heis. the security of cloud computing is a major factor that deters its adoption, as it plays a crucial role in protecting an organization's information and data. this study emphasizes the significance of the organizational perspective in innovation adoption, particularly the importance of top management commitment (tmc). tornatzky also recognized tmc as a crucial factor in the innovation adoption process (tornatzky, 1990). hence, the involvement of the top managers increases the chances of technology adoption. hence, it is especially true for developing countries like nigeria, with limited resources. so, strategically using limited resources and adopting cloud-based e-learning will considerably improve the over-stretched learning systems. therefore, we posit that: https://doi.org/10.34010/injiiscom.v3i2.9563 tom and virgiyanti. intention to adopt cloud-based e-learning in nigerian …| 244 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 2.4. h4: top management commitment will positively influence the adoption of cloudbased e-learning in nigerian heis. the natural viewpoint has an impact on how innovation is adopted in companies. cloud computing can offer a competitive edge to nigerian heis, but external factors like government regulations, policies, and peer pressure from global competitors can either support or hinder the adoption of new technology for learning systems. competitive pressure (cp) refers to the level of competition that an organization faces from other institutions within the academia. competition can drive heis to innovate and enhance the quality of their systems. therefore, the adoption of cloud-based elearning can help nigerian heis to restructure and create new opportunities for higher education. 2.5. h5: competitive pressure will positively influence the adoption of cloud-based e-learning in nigerian heis. this research also makes use of the toe and doi theories to forecast nigeria's propensity to embrace cloud-based elearning. the study uses poll questions to determine the critical toe-doi factors, which are shown in fig. 1. a seven-point likert measure was used to capture every answer. 3. research methodology deductive research technique, which includes the creation of quantifiable research queries, was used in this work. the study's element of research is the ict directorates' senior managerial employees. an easy random selection method was used to choose the group of colleges. the two parts of the poll items used in the research were modified from earlier studies. the subjects' personal data was collected in the first segment, and acceptance concerns for cloud-based e-learning were addressed in the second. responses were gathered using a 7-point likert scale, and the surveys received a topic validity assessment. a total of 248 responses were obtained from the 454 questionnaires distributed to the target respondents. fig. 1. the theoretical framework and hypothesis of the study https://doi.org/10.34010/injiiscom.v3i2.9563 245 | international journal of informatics information system and computer engineering 3(2) (2022) 241-250 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 4. measurement model in order to evaluate different parts of the data, including the content validity, internal consistency dependability, convergent validity, and discriminant validity, it is essential to evaluate the measurement model. these assessments assist in ensuring that the data gathered is precise and trustworthy as well as that the poll queries used are tracking the intended outcomes (henseler et al., 2009; hair et al., 2011; hair et al., 2014). 4.1. individual item reliability and internal consistency the internal consistency is a measurement of how closely various scale elements reflect the same fundamental entity. (bijttebier et al., 2000; sun et al., 2007). in most cases, the cronbach's alpha statistic is used to assess this. composite reliability factors are frequently used to calculate an instrument's internal consistency reliability. (bacon et al., 1995; mccrae et al., 2011; peterson & kim, 2013). in order to ensure the truth and dependability of a measurement tool, it is crucial to evaluate its internal coherence. 4.2. discriminant validity the degree to which two conceptually comparable conceptions differ from one another is referred to as discriminant validity (hair et al., 2010). it is assessed using the fornell and larcker-proposed average variance extracted (ave) technique (fornell & larcker, 1981). the common variation between the components should be less than the ave numbers. the entities are shown to be considerably distinct from one another if the ave values are greater than the common variation. the research data satisfies the requirement for discriminant validity, as shown in table 2. the htmt (heterotrait-monotrait ratio of correlations) is an additional method used to assess the discriminant validity of the data. table 3 shows the results of this analysis, indicating that the htmt is also not a concern, meaning that no items needed to be deleted in this phase. table 1. construct reliability and validity factors cronbach's alpha rho_a composite reliability ave com 0.843 0.858 0.883 0.557 cp 0.899 0.913 0.92 0.624 int 0.791 0.801 0.877 0.703 ra 0.816 0.843 0.86 0.508 sec 0.828 0.859 0.883 0.655 tmc 0.886 0.887 0.914 0.639 https://doi.org/10.34010/injiiscom.v3i2.9563 tom and virgiyanti. intention to adopt cloud-based e-learning in nigerian …| 246 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 table 2. fornell-lacker criterion variables com cp int ra sec tmc com .746 cp .576 .79 int .422 .564 .839 ra .399 .344 .44 .712 sec .466 .435 .402 .462 .809 tmc .569 .599 .438 .398 .623 .799 table 3. heterotrait-monotrait ratio (htmt) factors com cp int ra sec tmc com cp 0.661 int 0.475 0.633 ra 0.425 0.365 0.501 sec 0.53 0.48 0.471 0.515 tmc 0.653 0.659 0.499 0.417 0.704 4.3. assessment of structural model the structural model in this research was assessed using pls-sem (partial least squares structural equation modeling), as suggested by the conceptual model, which contained five assumptions (see fig. 2) (hair et al., 2017). following the recommendations made by hair et al., the importance of route coefficients was assessed using the conventional bootstrapping technique with 5,000 data. (2010, 2014, 2017). the normalcy of the data was estimated using the bootstrap findings. the structural model is shown in table 4, and the mediator variable is government support. according to the study's findings, there is a connection between competitive pressure (cp) and the desire of nigerian heis to embrace cloud-based e-learning. compatibility, security, and top management commitment, however, did not significantly influence whether cloud-based e-learning was adopted in heis. relative advantage (ra) and the desire to implement a cloud-based elearning system in nigeria were found to be significantly correlated in the research. in conclusion, these results shed important light on the importance of nigerian heis' plans to implement cloudbased e-learning technology. table 5 presents the findings. https://doi.org/10.34010/injiiscom.v3i2.9563 247 | international journal of informatics information system and computer engineering 3(2) (2022) 241-250 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 table 4. structural model with government support as the moderator variable factors sample mean (m) t statistics (|o/stdev|) p values com -> int 0.048 0.564 0.286 cp -> int 0.398 5.157 0.000 ra -> int 0.243 3.833 0.000 sec -> int 0.085 1.012 0.156 tmc -> int 0.029 0.306 0.380 fig. 2. the structural model table 5. bootstrapping findings hypothesis result h1 there will be a positive relationship between relative advantage (ra) and intention to adopt cloud-based e-learning. supported h2 there will be a positive relationship between compatibility (com) and intention to adopt cloud-based e-learning. not supported h3 there will be a positive relationship between security (sec) and intention to adopt cloud-based e-learning. not supported h4 there will be a positive relationship between top management commitment (tmc) and intention to adopt cloud-based e-learning. not supported h5 there will be a positive relationship between competitive pressure (cp) and intention to adopt cloud-based e-learning. supported 5. discussion this research set out to determine how eager nigerian hei administration was to implement cloud-based e-learning within their organizations. the research suggested an adaptation model based on the technology-organizationenvironment (toe) theory and other pertinent environmental factors to accomplish this objective. studies on cloud computing usage in heis around the globe are numerous, and the technology has many benefits over https://doi.org/10.34010/injiiscom.v3i2.9563 tom and virgiyanti. intention to adopt cloud-based e-learning in nigerian …| 248 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 conventional e-learning platforms. therefore, incorporating cloud computing into nigerian heis is crucial. the results of the research show that the toe theory elements are extremely important for comprehending and affecting the desire to embrace cloudbased e-learning. relative advantage, a technical component, had a substantial effect on the desire to embrace cloud-based elearning, according to the structural equation modeling (sem) study, with βvalue = 0.062, t-value = 3.833, and p-value = 0.000. the upper management views the impact of this technology as vital, in line with roger's diffusion of innovation (doi) theory. (rogers, 2003). according to them, adopting cloud-based elearning will enhance the standard of procedures and employee performance in their institutions, which is consistent with tashkandi and al-jabri's results. (tashkandi & al-jabri, 2015). on the other hand, with a p-value of 0.286 and a β-value of 0.072, compatibility is not a major factor. this suggests that heis in nigeria are still adopting cloud-based elearning in its early phases and have not yet concentrated on interoperability with their existing operational systems (rogers, 2003). and others, however, contend that compatibility is a crucial aspect of technology usage in heis (hiran & henten, 2020). the study's findings also demonstrated that top management commitment (tmc), as an organizational component, had little bearing on nigerian heis' uptake of cloud-based e-learning. this result emphasizes the difficulties and opposition that heis encounter when implementing innovation and change, and it implies that senior administrators should be more knowledgeable about cloud technology. on the other hand, the security factor was also found to be statistically insignificant. despite being a major concern for many heis, managers are still cautious about cloud computing due to issues such as data ownership, location, privacy, confidentiality, and data availability. data ownership, in particular, remains a major challenge for cloud computing as the policies of the country where the data centres are located may require them to make the data available to them. 6. conclusion the purpose of this research was to investigate how university senior management in northern nigeria perceived the usage of cloud-based elearning and the variables that might have an effect on that goal. based on the work of tornatzky and fleischer, the technology-organization-environment theory was applied and extended to include additional pertinent factors for the research (tornatzky & fleischer's, 1990). the pls-sem method was used to evaluate the interviewees' data, and the results showed that relative advantage had a substantial impact on the management's decision to implement cloud-based e-learning in nigerian heis. the desire to embrace cloud-based elearning was also found to be significantly impacted by competitive pressure. however, it was discovered that crucial elements like compatibility, security, and top management commitment were negligible in nigeria. to enhance their competitiveness with global peers, nigerian heis must gain an understanding of cloud computing and make informed decisions about adopting https://doi.org/10.34010/injiiscom.v3i2.9563 249 | international journal of informatics information system and computer engineering 3(2) (2022) 241-250 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 cloud-based e-learning. financial support is crucial for implementing this technology, which is widely used by developed nations to provide efficient and effective learning experiences. although more research is required, this survey provides insightful information about the desire to implement cloudbased e-learning in nigerian heis. future studies could concentrate on cloud-based e-learning in nigeria and make use of mixed-methods or qualitative techniques to give participants' views of implementing this technology in nigerian heis a more thorough grasp. due to financial and organizational limitations, this research was only able to encompass a small number of colleges in a particular area of nigeria. future studies could use a larger sample size and responses from staff and other managerial levels to get a more thorough grasp of the usage of cloud-based elearning in nigerian heis. future study could use online surveys and conversations to collect information on the crucial factors influencing the uptake of cloud-based e-learning in nigerian heis. references asiyai, r. i. (2013). challenges of quality in higher education in nigeria in the 21st century. international journal of educational planning & administration, 3(2), 159– 172. bacon, d. r., sauer, p. l., & young, m. (1995). composite reliability in structural equations modeling. educational and psychological measurement, 55(3), 394–406. https://doi.org/10.1177/0013164495055003003 bijttebier, p., delva, d., vanoost, s., bobbaers, h., lauwers, p., & vertommen, h. (2000). reliability and validity of the critical care family needs inventory in a dutch-speaking belgian sample. heart & lung, 29(4), 278–286. https://doi.org/10.1067/mhl.2000.107918 edomwonyi, j., & osarumwense, r. (2017). business education in nigeria: issues, challenges and way forward for national development. journal of collaborative research and development (jcrd), 5(1), 1–25. fornell, c., & larcker, d. f. (1981). evaluating structural equation models with unobservable variables and measurement error. journal of marketing research, 18(1), 39. https://doi.org/10.2307/3151312 hair, j. f., black, w. c., babin, b. j., anderson, r. e., & tatham, r. l. (2014). multivariate data analysis (7th edition). pearson education limited. www.pearsoned.co.uk hair, j. f. j., black, w. c., babin, b. j., & anderson, r. e. (2010). multivariate data analysis (seventh ed). pearson. https://www.pearson.com/us/highereducation/program/hair-multivariate-data-analysis-7thedition/pgm263675.html hair, joe f., ringle, c. m., & sarstedt, m. (2011). pls-sem: indeed a silver bullet. the journal of marketing theory and practice, 19(2), 139–152. https://doi.org/10.2753/mtp1069-6679190202 https://doi.org/10.34010/injiiscom.v3i2.9563 tom and virgiyanti. intention to adopt cloud-based e-learning in nigerian …| 250 doi: https://doi.org/10.34010/injiiscom.v3i2.9563 p-issn 2810-0670 e-issn 2775-5584 hair, joseph f., hult, jr., g. t. m., ringle, c., & sarstedt, m. (2014). a primer on partial least squares structural equation modeling (pls-sem). sage publications,. hair, joseph f., hult, g. t. m., ringle, c., & sarstedt, m. (2017). a primer on partial least squares structural equation modeling (pls-sem) (second edi). https://uk.sagepub.com/en-gb/asi/a-primer-on-partial-least-squaresstructural-equation-modeling-pls-sem/book244583 henseler, j., ringle, c. m., & sinkovics, r. r. (2009). the use of partial least squares path modeling in international marketing. emerald group publishing limited, 20, 277–319. https://doi.org/10.1108/s1474-7979(2009)0000020014 hiran, k. k., & henten, a. (2020). an integrated toe-doi framework for cloud computing adoption in higher education: the case of sub-saharan africa, ethiopia. in soft computing: theories and applications (pp. 1281–1290). springer. mccrae, r. r., kurtz, j. e., yamagata, s., & terracciano, a. (2011). internal consistency, retest reliability, and their implications for personality scale validity. personality and social psychology review, 15(1), 28–50. https://doi.org/10.1177/1088868310366253 peterson, r. a., & kim, y. (2013). on the relationship between coefficient alpha and composite reliability. journal of applied psychology, 98(1), 194–198. https://doi.org/10.1037/a0030767 rogers, e. m. (2003). diffusion of innovations (5th ed., vol. 1, issue 38). sun, w., chou, c.-p., stacy, a. w., ma, h., unger, j., & gallaher, p. (2007). sas and spss macros to calculate standardized cronbach’s alpha using the upper bound of the phi coefficient for dichotomous items. behavior research methods, 39(1), 71– 81. https://doi.org/10.3758/bf03192845 tashkandi, a. n., & al-jabri, i. m. (2015). cloud computing adoption by higher education institutions in saudi arabia: an exploratory study. cluster computing, 18(4), 1527–1537. https://doi.org/10.1007/s10586-015-0490-4 tom, a. m., virgiyanti, w., & osman, w. r. s. (2019). the impact of government support on the adoption of iaasbel by university’s top management. 2019 international conference on data and software engineering (icodse), 1–6. tom, a. m., virgiyanti, w., & rozaini, w. (2019). understanding the determinants of infrastructure-as-a service-based e-learning adoption using an integrated toe-doi model: a nigerian perspective. 2019 6th international conference on research and innovation in information systems (icriis), 1–6. https://doi.org/10.1109/icriis48246.2019.9073418 tornatzky, l. g., & fleischer, m. (1990). the processes of technological innovation. the journal of technology transfer, 16(1), 45–46. https://doi.org/10.1007/bf02371446 world bank. (2013). smarter education systems for brighter futures. https://doi.org/10.34010/injiiscom.v3i2.9563 49 | international journal of informatics information system and computer engineering 4(1) (2023) 49-60 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 a computational bibliometric analysis of game advertising using vosviewer yogie rinaldy ginting department of mechanical engineering, school of mechanical and civil engineering, curtin university, australia *corresponding email: y.ginting@curtin.edu.au 1. introduction a computational bibliometric analysis of game advetising using vosviewer. advetising games are attitudes related to a person's financial problems, where responses to a statement or opinion can be used as a useful measurement variable to define advetising games as a state of mind, opinions and judgments about finances. the game consists of a set of rules that build competing situations from two to several people or groups by choosing strategies built to maximize a b s t r a c t s a r t i c l e i n f o the purpose of this study is to perform a computational bibliometric analysis of the term “game advertising” by combining mapping analysis using vosviewer publish or perish software, and google schoolar. the method used is a bibliometric and descriptive quantitative approach. the data obtained is a search result based on the keyword "game advertising" on google scholar. the search results show 989 articles published from 2017 to 2022 decreasing every year except in 2021. this can be proven in 2017 with 232 articles, in 2018 it decreased to 210 articles, in 2019 it decreased again to 199 articles. in 2020 it decreased to 136 articles, except for in 2021 there was a less significant increase to 137 articles, and in 2022 research on game advertising decreased dramatically, the number of publications to 40. the conclusion of this study shows the importance of conducting bibliometric analysis, especially in advertising games field. this research is expected to be a reference for further research in determining and analyzing the research theme. article history: submitted/received 03 nov 2022 first revised 20 feb 2023 accepted 25 mar 2023 first available online 12 apr 2023 publication date 01 jun 2023 aug 2018 __________________ keywords: bibliometrics, game advertising, data analysis, vosviewer international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(1) (2023) 49-60 https://doi.org/10.34010/injiiscom.v4i1.9571 mailto:y.ginting@curtin.edu.au yogie rinaldy ginting. a computational bibliometric analysis of game advertising…| 50 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 their own winnings or to minimize your opponent's winnings. the rules specify the possibility of action for each player, a certain amount of information received by each player as the progress of play, and a certain number of wins or losses in various situations (von eumann & morgenstern, 2007). the west kutai regency government aims to enhance the effectiveness and efficiency of public services by implementing e-government internally within government organizations. one aspect of egovernment that must be prioritized is government to citizen or government to customer (g2c) services, where the government provides public service information through information technology. online services can significantly reduce administrative, relational, and interaction costs compared to manual services, benefiting both the government and its stakeholders. additionally, there are technologies that can simplify administrative processes and reduce bureaucracy, thereby creating a positive business environment. however, based on direct observations including interviews and discussions with kesbangpol of west kutai regency, it is evident that the implementation of egovernment is still in the preparation stage and has not yet reached the emerging level or the lowest level of adoption. promotion is anything that is done to help sell a product or service at each point of the sales network, from the presentation materials a salesperson uses when making offers to commercial broadcasts, television, or newspaper advertisements that try to lure customers into getting favorable impressions of what is advertised (goutama, ae., 2018). advertising games (adv-games) are video games that include promotional content as a marketing technique, either overtly or implicitly. the brand is integrated into the story, mission, and other game activities in an adv-game. these adv-games can be observed in a variety of game components, including characters with specific brands, gameplay that demonstrates the characteristics of specific products, advertising banners in a game segment, and other elements (aulia, et al, 2014). in general, advanced games have been around for a long time, yet they continue to evolve and adapt. traditional games like monopoly, snakes and ladders, and chess may all be turned into adv-games by adding a hint of brand awareness. following technical advancements, advgames have evolved into digital media, and various adv-games applications are now available on smartphone media (terlutter and capella, 2013). advertising games (adv-games) can be an opportunity for business people to improve their marketing tools. to determine the great opportunity about improving marketing advertising through certain video games, research on each video game, including advertising games in improving marketing tools, research using bibliometrics should also be carried out. this is because bibliometrics can be used to categorize certain topics in bibliographical form as well as to generate representative summaries of selected topics. vosviewer is a software tool that facilitates bibliometric analysis, offering the capability to generate, visualize, and analyze bibliometric maps. with vosviewer different types of bibliometric network data can be evaluated, such as connections between https://doi.org/10.34010/injiiscom.v4i1.9571 51 | international journal of informatics information system and computer engineering 4(1) (2023) 49-60 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 journal publications or citations, associations between scientific terminology, and collaborative relationships among researchers. building upon previous research, this study aims to conduct bibliometric analysis on the topic of game advertising by utilizing vosviewer software for mapping analysis. the research methodology involves quantitative bibliometric analysis and a descriptive approach. data is collected from searches and processing using the keyword "game advertising" on platforms such as google scholar and publish or perish (kania and sabariah, 2013). 2. method the research methodology employed in this study involves bibliometric, descriptive, and quantitative approaches. information was gathered from various published journals indexed by google scholar. additionally, a literature review was conducted on the research topic "gaming advertising" using the publish or perish software, chosen specifically for identifying bibliometric data (arwendria, a., 2021). furthermore, data obtained from publish or perish was saved in *ris format and imported into vosviewer software for analysis. publish or perish 8 and vosviewer 1.6.17 were utilized as the software tools for data collection in this study. in this research, relevant materials related to game advertising were reviewed and selected. a total of 989 research articles on game advertising published between 2017 and 2022 were obtained. these articles were saved in *.ris format for further analysis. vosviewer software was then used to generate visualizations and analyze trends using bibliometric maps. the article data from the prepared database sources were mapped using three types of visualizations in vosviewer software: network visualization, overlay visualization, and density visualization. additionally, terms included in the vosviewer mapping visualization were filtered to refine the analysis (kurnia, s., 2021). 3. results and discussion 3.1. advancements in the field of game advertising research research on the development climate in the field of game advertising, illustrating the development of research in the field of game advertising from 2017 to 2022 in fig. 1. fig. 1. level of research development on game advertising figure 1 shows that research on game advertising decreases every year, starting from 2017 to 2022 except for 2021. this can be proven in 2017 with 232 articles, in 2018 it decreased to 210 articles, in 2019 it decreased again to 199 articles, in 2020 it decreased to 136 articles, except in 2021 there was a less significant increase to 137 articles, and in 2022 research on game advertising decreased dramatically, the number of publications to 40. after conducting a search using the publish or perish software, we identified 989 articles https://doi.org/10.34010/injiiscom.v4i1.9571 yogie rinaldy ginting. a computational bibliometric analysis of game advertising…| 52 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 that are relevant to the research topic. from this dataset, we further filtered and selected the 20 articles with the highest number of citations, sourced from 20 distinct journals and books (see table 1). table 1. article data in the field of game advertising no authors title year cites refs 1. herbert., et a l. ma rketing: grundla gen ma rktorientierter unternehmensführung konzepte – instrumente – pra xisbeispiele 2018 7173 (herbert, et a l, 2018) 2. silver., et a l. ma stering the ga me of go without huma n knowledge 2017 7044 (silver, et a l, 2017) 3. leiss., et a l. socia l communica tion in a dvertising 2018 2195 (leiss, et a l, 2018) 4. silver., et a l. a genera l reinforcement lea rning a lgorithm tha t ma ster's chess, shogi, a nd go through self -pla y 2018 2194 (silver, et a l, 2018) 5. j koivisto., & j ha ma ri. the rise of motiva tiona l informa tion systems: a review of ga mifica tion resea rch 2019 720 (j koivisto & j ha ma ri, 2019) 6. k huota ri., & j ha ma ri. a definition for ga mifica tion: a nchoring ga mifica tion in the service ma rketing litera ture 2017 685 (k huota ri & j ha ma ri, 2017) 7. eva ns., et a l. disclosing insta gra m influencer a dvertising: the effects of disclosure la ngua ge on a dvertising recognition, a ttitudes, a nd beha viora l intent 2017 611 (eva ns, et a l, 2017) 8. lee., et a l. advertising content a nd consumer enga gement on socia l media : evidence from fa cebook 2018 605 (lee, et a l, 2018) 9. s hea lth. method for socia l networking intera ctions using online consumer browsing beha vior, buying pa tterns, a dvertisements a nd a ffilia te a dvertising, for promotions, online 2019 590 (s hea lth, 2019) 10. db nieborg ., & t poell. the pla tformiza tion of cultura l production: theorizing the contingent cultura l commodities 2018 503 (db nieborg ., & t poell, 2018) 11. wa ng., et a l. irga n : a minima x ga me for unifying genera tive a nd discrimina tive informa tion retrieva l models 2017 460 (wa ng, et a l, 2017) 12. steenka mp., et a l. competitive rea ctions to a dvertising a nd promotion a tta cks 2018 367 (steenka mp, et a l, 2018) 13. m meeker., & l wu. internet trends 2018 2019 321 (m meeker, & l wu, 2019) 14. j roozenbeek ., & s va n der linden. fa ke news ga me conferences psychologica l resista nce a ga inst online misinforma tion 2019 244 (j roozenbeek & s va n der linden, 2019) 15. s de freita s. are ga mes effective lea rning tools? a review of educa tiona l ga mes 2018 243 (s de freita s, 2018) https://doi.org/10.34010/injiiscom.v4i1.9571 53 | international journal of informatics information system and computer engineering 4(1) (2023) 49-60 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 table 1 (continue). article data in the field of game advertising no authors title year cites refs 16. s da hl. socia l media ma rketing: theories a nd a pplica tions 2021 213 (s da hl, 2021) 17. bt sha piro. positive spillovers a nd free riding in a dvertising of prescription pha rma ceutica ls: the ca se of a ntidepressa nts 2018 188 (bt sha piro, 2018) 18. wa ng., et a l. displa y a dvertising with rea l-time bidding (rtb) a nd beha viora l ta rgeting 2017 157 (wa ng, et a l, 2017) 19. smith., et a l. food ma rketing influences children's a ttitudes, preferences a nd consumption: a systema tic critica l review 2019 152 (smith, et a l, 2019) in table 1 there are 19 articles that match the criteria research. of the 19 selected articles, showing that highest quote related to game advertising research is 7173, while with the lowest citation is 152. that in table 1, it shows that in 2017 and 2022, each has articles with quotes highest. in 2017-2022, the most articles quoted is 7173 articles. temporary that, in 2021, a lot of articles quoted are 213 articles. year with quote the most is in 2018 as many as 7173 articles. 3.2. visualization game advertising topic area using vosviewer the visualization of the game advertising research area was carried out using vosviewer software, with a minimum requirement of 3 relationships, with 2 terms set by al husaeni and nandiyanto (utami and karlina, 2022). as a result, a total of 26 items were obtained and clustered into 4 groups based on the analysis of mapping visualization, according to the study on game advertising climate, namely: (i) cluster 1 has 9 items, those 9 items elements, engagement, game elements, gamers, gamified systems, market research, marketing communication, motivation, and tourism marketing (see figure 2). (ii) cluster 2 has 8 items, those 8 items are application, case study, computer game, development, digital marketing, education, marketing strategy, and target market (see figure 3). (iii) cluster 3 has 5 items, those 5 items are gamification, gamified application, interactivity, and service marketing perspective (see figure 4). (iv) cluster 4 has 4 items, those 4 items are analysis, game advertising, marketing management, and research directions (see figure 5). cluster 1 is indicated by color red, cluster 2 is shown in color green, cluster 3 is shown in blue old, and cluster 4 is shown in color yellow. https://doi.org/10.34010/injiiscom.v4i1.9571 yogie rinaldy ginting. a computational bibliometric analysis of game advertising…| 54 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 fig. 2. cluster 1 visualization advertising game network fig. 3. cluster 2 visualization advertising game network https://doi.org/10.34010/injiiscom.v4i1.9571 55 | international journal of informatics information system and computer engineering 4(1) (2023) 49-60 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 fig. 4. cluster 3 visualization advertising game network fig. 5. cluster 4 visualization advertising game network 3.3. network visualization game advertising topic area using vosviewer in vosviewer software, mapping every term divided becomes three type, the first is visualization network. visualization network that is connection among thing on the map. existing relationship in network visualization shown in network or the line that goes from one to one other things (see figure 6). https://doi.org/10.34010/injiiscom.v4i1.9571 yogie rinaldy ginting. a computational bibliometric analysis of game advertising…| 56 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 fig. 6. visualization advertising game network visualization network from the "advertising game" gained using the vosviewer software is displayed in fig. 6. each cluster where in each each each field or investigated issue is represented in fig. 6. the game advertising climate alone is included in cluster 4 with a total strength of 40 and an occurrence of 43, as indicated in figure 6 above. advetising game climate connected to cluster 1, among them that is terms game element, marketing communication, and gamer, in k laster 2, among others that is term application, marketing strategy, computer games, and finally in cluster 3 of them namely with the term gamify, interactivity. 3.4. overlay visualization of game advetising topic area using vosviewer b second visualization network, in vosviewer software provide visualization mapping in overlay shape. mapping in shape overlay visualization focuses on novelty something term in research. novelty term or thing in research related to the climate of advertising games shown in figure 7. https://doi.org/10.34010/injiiscom.v4i1.9571 57 | international journal of informatics information system and computer engineering 4(1) (2023) 49-60 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 fig. 7. overlay game advertising visualization in the depiction thing or term type overlay visualization, can see how much popular from every year. on visualization overlay, different colors showing extension term in something period certain. in research this, we use year 2017 to year 2022. more colors dark approach purple meaning that in something study about one thing or term done more close to 2017. meanwhile, the color is more light approach yellow is existing term in study latest. 3.5. density visualization of game advertising third that is the last mapping depiction in the vosviewer software is density visualization. visualization of density on financial attitude is shown in fig. 8. the mapping type used in this study involves the use of colors to indicate the popularity of a term. if the color of a term is lighter, it signifies that research on that term is becoming more popular. conversely, if the color is darker or faded, it indicates that research on that term is decreasing in frequency. in figure 8, it can be observed that there are some terms depicted in yellow color with relatively larger diameters. these terms called emission, game advertising, analysis, application, and engagement. https://doi.org/10.34010/injiiscom.v4i1.9571 yogie rinaldy ginting. a computational bibliometric analysis of game advertising…| 58 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 fig. 8. visualization density advertising game network visualization density about climate advertising game research is in the picture above, which means that on the map density showing results analysis use all article regarding game advertising in 2017-2022. in the visualization in figure 8 shows pattern colored yellow where is getting yellow color, more keyword thick and getting the diameter of the circle, which means that they appear more dominant and if color on map fade or blend with the background behind colored green, shows that keyword the appear more rarely 4. conclusion study this aims to examine, analyze bibliometric literature on game advertising. the keyword "game advertising " is used to obtain data, which is based on a topic area that contains keywords, abstracts, and titles. after processing and filtering the data, obtained 989 relevant articles. device soft vosviewer is used to generate mapping data. mapping data poured to in visualization grid, overlay, and density. based on results in mapping and analysis use vosviewer, obtained that study regarding financial management with the term game advertising in 2017-2022 decreased from every year to year. in research this, using method bibliometrics to identify theme main in every field studies before, because important to assess novelty in future research. https://doi.org/10.34010/injiiscom.v4i1.9571 59 | international journal of informatics information system and computer engineering 4(1) (2023) 49-60 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 references arwendria , a. (2021). publish or perish: bibliometric analysis of the literature on covid-19 on the google scholarship data base for 2019-2021. alma'arif : science islamic libraries and information , 1(1), 1-12. aulia, eg.; nurkertamanda, d.; and budiawan , w. (2014). analysis of the effectiveness of dynamic in-game advertising on android games with eyetracking method. industrial engineering online journal, 3(1). dahl, s. (2021). social media marketing: theories and applications. social media marketing, 1-100. de freitas, s. (2018). are games effective learning tools? a review of educational games. journal of educational technology and society, 21(2), 74-84. evans, nj; phua, j.; lim, j.; and jun, h. (2017). disclosing instagram influencer advertising: the effects of disclosure language on advertising recognition, attitudes, and behavioral intent. journal of interactive advertising, 17(2), 138149. goutama , ae (2018). games advertising vs tv advertising which one is more effective in building brands? stei journal of economics, 27(01), 1-9. heath, s. (2019). u.s. patent no. 10,217,117. washington, dc: us patent and trademark office. huotari, k.; and hamari, j. (2017). a definition for gamification: anchoring gamification in the service marketing literature. electronic markets, 27(1), 2131. kania, mbddm, and sabariah, a. (2013). analysis quality software against _ system unikom information . magazine unikom science. koivisto, j.; and hamari, j. (2019). the rise of motivational information systems: a review of gamification research. international journal of information management, 45, 191-210. kurnia, s. (2021). science, technology, engineering, art and mathematics (steam) in science education: analysis bibliometrics and mapping literature study use vosviewer software ( doctoral dissertation, uin raden intan lampung). lee, d.; hosanagar , k.; and nair, hs (2018). advertising content and consumer engagement on social media: evidence from facebook. management science, 64(11), 5105-5131. leiss, w.; kline, s.; jhally , s.; botterill, j.; and asquith, k. (2018). social communication in advertising. routledge. meeker, m., and wu, l. (2018). internet trends 2018. meffert, h.; burmann, c.; kirchgeorg , m.; and eisenbeiß , m. (2018). marketing: grundlagen marktorientierter unternehmensführung konzepte – instrumente – praxisbeispiele. springer-verlag. https://doi.org/10.34010/injiiscom.v4i1.9571 yogie rinaldy ginting. a computational bibliometric analysis of game advertising…| 60 doi: https://doi.org/10.34010/injiiscom.v4i1.9571 p-issn 2810-0670 e-issn 2775-5584 roozenbeek , j.; and van der linden, s. (2019). fake news game conferences psychological resistance against online misinformation. palgrave communications, 5(1), 1-10. shapiro, bt (2018). positive spillovers and free riding in advertising of prescription pharmaceuticals: the case of antidepressants. journal of political economy, 126(1), 381-437. shapiro, bt (2018). positive spillovers and free riding in advertising of prescription pharmaceuticals: the case of antidepressants. journal of political economy, 126(1), 381-437. silver, d., hubert, t., schrittwieser, j., antonoglou, i., lai, m., guez, a., ... & hassabis, d. (2018). a general reinforcement learning algorithm that master’s chess, shogi, and go through self-play. science, 362(6419), 1140-1144. silver, d.; schrittwieser , j.; simonyan, k.; antonoglou , i.; huang, a.; guez, a.; ... and hassabis, d. (2017). mastering the game of go without human knowledge. nature, 550(7676), 354-359. smith, r.; kelly, b.; yeatman, h.; and boyland, e. (2019). food marketing influences children's attitudes, preferences and consumption: a systemati c critical review. nutrients, 11(4), 875. steenkamp, jbe.; nijs, vr.; hanssens , dm.; and dekimpe , mg (2018). competitive reactions to advertising and promotion attacks. in long-term impact of marketing: a compendium 325-372. terlutter, r.; and capella, ml (2013). the gamification of advertising: analysis and research directions of in-game advertising, advergames, and advertising in social network games. journal of advertising, 42(2-3), 95-112. utami , sb; and karlina, n. (2022). bibliometric analysis: development of research and publications on program coordination using vosviewer. journal of cultural libraries , 9 (1), 1-8. von neumann, j.; & morgenstern, o. (2007). theory of games and economic behavior. in theory of games and economic behavior. princeton university press. wang, j.; yu, l.; zhang, w.; gong, y.; xu, y.; wang, b.; ... and zhang, d. (2017, august). irgan : a minimax game for unifying generative and discriminative information retrieval models. in proceedings of the 40th international acm sigir conference on research and development in information retrieval, 515-524. wang, j.; zhang, w.; and yuan, s. (2017). display advertising with real-time bidding (rtb) and behavioral targeting. foundations and trends® in information retrieval, 11(4-5), 297-435. https://doi.org/10.34010/injiiscom.v4i1.9571 samuel w lusweti et al. impact of number of artificial ants in aco … | 124 impact of number of artificial ants in aco on network convergence time: a survey samuel w lusweti1*, collins o odoyo, dorothy a rambim department of information technology, school of computing and informatics, masinde muliro university of science and technology, kenya *corresponding email: lusweti015@gmail.com a b s t r a c t s a r t i c l e i n f o due to the dynamic nature of computer networks today, there is need to make the networks self-organized. selforganization can be achieved by applying intelligent systems in the networks to improve convergence time. bio-inspired algorithms that imitate real ant foraging behaviour of natural ants have been seen to be more successful when applied to computer networks to make the networks self-organized. in this paper, we studied how ant colony optimization (aco) has been applied in the networks as a bio-inspired algorithm and its challenges. we identified the number of ants as a drawback to guide this research. we retrieved a number of studies carried out on the influence of ant density on optimum deviation, number of iterations and optimization time. we found that even though some researches pointed out that the numbers of ants had no effect on algorithm performance, many others showed that indeed the number of ants which is a parameter to be set on the algorithm significantly affect its performance. to help bridge the gap on whether or not the number of ants were significant, we gave our recommendations based on the results from various studies in the conclusion section of this paper. article history: received 25 may 2022 revised 30 may 2022 accepted 10 june 2022 available online 26 june 2022 aug 2018 __________________ keywords: convergence time, ant colony optimization, artificial ants, networks, parameter 1. introduction a family of optimization techniques that have been applied as combinatorial problem-solving techniques form the widely known metaheuristics. over the years, metaheuristics have been applied in many fields to solve complex problems (liao et al., 2014). the aco algorithm is international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 124-135 125 |international journal of informatics information system and computer engineering 3(1) (2022) 124-135 one example of these metaheuristics. others include particle swarm optimization, bee colony optimization, bat optimization among others. aco was introduced in the early years of the 1990s (blum, 2005) and has for many years been applied in various fields to tackle complex optimization issues that may be solved using basic methods. for example, the algorithm has been successfully applied in data compression, gaming theory, feature selection, dispatch problems, parameter estimation in dynamic systems, satellite control, job scheduling problems, congestion control, social graph mining, medicine for decision making, and target tracking (kar, 2016). aco has also been employed in computer networks to determine the shortest possible route for sending data from basis to destination, similar to how ants forage (adiga et al., 2013), therefore becoming the lowest-cost route between any two connected nodes. this has increased network functioning while decreasing the latency that packets suffer in the absence of the aco algorithm, when attempting to reach their target because computer networks are important to the running of any institution, they must be active at all times, regardless of the obstacles they may encounter. as a result, technical adjustments in their upkeep are required. aco was employed in this example to make the network self-regulating and sustainable, so that any issues discovered in the network could be addressed by the network itself. some academics have proposed using aco to prevent the need for infrastructure such as nodes and switches, which might fail and cause the communication process to fail (dressler et al., 2010). 2. methods and materials in this paper we adopted both qualitative research methods. secondary data of 12 experiments published online was retrieved whereby results from various studies were compared and used to do the analysis using thematic analysis. 3. application of aco in computer networks ant colony optimization algorithm is a class of swarm intelligent systems that are applied in solving np-hard problems. swarm-intelligent systems have not been fully explored in literature till today (gustov & levina, 2021), but a lot is being done in aco. this is because among the many swarms intelligent algorithms available, aco is the most studied. foraging ants in a true ant colony system produce pheromone trails along the pathways they traverse, which others may follow to the food source. an ant in the aco algorithm is a mobile agent capable of dealing with computer network difficulties such as congestion and packet routing. this is made possible by the continuous and consistent modification of the routing tables by the agents in response to any congestion in the network (sim & sun, 2002). as the agents do this modification, they lay pheromone trails, which make clear routes between any nodes on the system. these pheromone trails acting as stigmergy would be handy in helping the future ants make a routing decision (kamali & opatrny, 2007). other biosamuel w lusweti et al. impact of number of artificial ants in aco … | 126 inspired algorithms that have been successfully applied in computer networks for adaptive routing include the artificial plant optimization (apo) algorithm for the implementation of the telecom sensor network, artificial neural networks (ann) for switching networks, the genetic algorithm (ga) for network path routing and the leaping frog algorithm (lfa) for network designing and scaling problems (kar, 2016). however, aco has been the most often used and researched optimization algorithm since it has shown the highest performance and has been the most effective (caro & dorigo, 2004). as a result, this research concentrated on the design and use of aco in many sectors, and more especially on its application in computer networks, and how it may even be improved in terms of the ideal number of ants required to be employed for better performance in these networks. forward and backward ants are used in these algorithms (jacobsen et al., 2011). the proactive aco routing algorithm antnet has been used effectively in packet-switched networks (caro & dorigo, 1998). the antnet algorithm propels a forward ant from the nest or source node at regular intervals towards its objective of food. when a forwarding ant reaches at its destination, it uses the list to return to the nest or source and update the pheromone values deposited in the routes or connections. if aco is applied in an ideal network, the ants are translated into packets, and the routes they use become the links between the nodes on that very network. if we have redundant links between the devices on the network, then the packets are expected to go through the shortest route to the destination if they are well optimized. 4. drawbacks and variations of aco the aco has shown certain benefits, such as positive feedback for quick solutions, dynamic applications that adjust to changes such as additional distances, and intrinsic parallelism. it has, however, shown several limitations, such as probability distribution changes due to iterations and convergence time (selvi & umarani, 2010). more precisely, anthocnet's disadvantage is the quantity of routing messages routing that must be delivered in the network before the formation of routes to the destination, the downside of antnet is the time necessary to build a route system between any two nodes in the network. this is referred to as the convergence time (selvi & umarani, 2010). last but not least the stagnation of ants in the working process of the algorithm (caro & dorigo, 1998) is also another problem common in aco algorithms. parameter setting of a basic ant colony algorithm is mainly the cause of these variations and is still under experimental stage till today (wei, 2014). 4.1 aco algorithm parameter setting the following are the parameters under consideration (wei, 2014). m number of ants. α pheromone relative importance. β relative importance of heuristic factor. 127 |international journal of informatics information system and computer engineering 3(1) (2022) 124-135 ρpheromone evaporation coefficient while (1-ρ) indicates the pheromone persistence factor. q amount of pheromone released by ants the following is the general formula used in the algorithm with the above parameters. the ants would independently select the next town or city to tour at time t, hence the probability of ant k to move from city i to j is given by (caro & dorigo, 1998): after all the ants have toured all the cities in the search space, each of the paths is updated according to eq. (2) below. were, and, this study compares various researches on the impact of number of ants (m) shown in eq. (3) on optimal solution and convergence time of the algorithm. 5. the effect of the number of ants (m) on aco optimization according to (colorni et al., 1991) the number of ants is among the controllable parameters affecting the performance of aco. we examine at 12 tests from three research to see how the number of ants in the aco algorithm affects its convergence. here, we look at whether or not an optimal solution is found and how long it takes for the solution to be found. 5.1. experiment 1 this experiment was done by aydin and yilmaz (sivagaminathan & ramakrishnan, 2007). the two presented an investigation into the number of ants used in aco in relation to the number of iterations, penalized objective function, and optimization time. for the purposes of this study, the results obtained from the number of iterations as well as time of optimization versus the number of ants are taken into account (see figs. 1). fig. 1. the average iteration number versus number of ants (sivagaminathan & ramakrishnan, 2007) fig. 2. the optimization time versus number of ants (sivagaminathan & ramakrishnan, 2007) samuel w lusweti et al. impact of number of artificial ants in aco … | 128 from fig. 1 above, as the number of ants rises, the number of iterations reduces. in contrast to fig. 2 above, the optimization time increases quickly when the population of ants grows. in terms of the precise number of ants required for the optimum solution, this research does not give a direct answer, as it only suggests that it is critical to identify the appropriate number of ants in order to get the best solution in the shortest amount of time. however, from the experimental results in fig. 1 and 2 above, it is obvious that the fewer the ants, the greater the number of iterations, and hence the shorter the optimization time. in other terms, when the number of ants rises, the number of iterations decreases yet the optimization time increases since numerous ants take longer to converge. however, this experiment was limited to small problems, and the exact number of optimal ants could not be established clearly. 5.2. experiment 2 we examine this experiment done by alobaedy et al (yilmaz & aydin). in this experiment, the researchers categorized optimization problems into small and medium scales using data sets of 50 and 100 cities respectively. they were able to measure the execution time, best solution, total number of new solutions obtained among other metrics. fig. 3 below shows the results obtained in terms of execution time against the number of ants for small scale problem of 50 cities. figure 3 shows that increasing the number of ants causes an increase in the execution time. fig. 3. execution time against the number of ants (50 cities) (yilmaz & aydin) when the experiment with 100 ants (medium) was run as shown below in fig. 4, a similar pattern was noted. however, due to the complexity of this issue, the execution time of the method increased in fig. 4. fig. 4. execution time against the number of ants (100 cities) (yilmaz & aydin) this research reveals that increasing the number of ants did not improve the algorithm's performance, but if the number is low, the performance was enhanced. nevertheless, the experiment was performed under two problem variations that is small and medium sized problems. for small sized problem the execution time was small and difficult to determine the exact number of ants needed to optimize the solution. whereas on the medium sized problem, the 129 |international journal of informatics information system and computer engineering 3(1) (2022) 124-135 execution time doubled and the number of optimum ants for that solution was found to be 16 out of 100. 5.3. experiment 3 in this experiment, christoffer and lars (alobaedy et al., 2017; stutzle & hoos, 2000; aydin) carried out comparisons on three aco variations model (rankedas eliteas, and mmas). the eliteas relies on specialist ants to work. these secondary ants, or specialized ants, are utilized to impose the elitist approach. the other ants called the normal ants work differently. the specialist ants multiply the pheromone on the best solution found by normal ants (petterson & lundell, 2018) making it stay longer without decomposing unlike the normal routes in other ant systems. the rankedas on the other hand also use specialist ants but these ants deposit pheromones on many good paths found instaed of depositing all pheromones on the best solution found (petterson & lundell, 2018). every route is graded by length, with the best-ranked route getting the most pheromones and the worstranked route receiving the fewest. lastly, the mmas has no specialist ants which means it only uses the normal ants. here, the pheromone deposited on a given path can never exit a maximum value or getlower than a given minimum value. this ensures that the pheromone level on a path does not get too low that the path is rendered unusable or the path should never be filled with the pheromone so much that it overshadows all the other routes (bullnheimer et al., 1997). this happens through smoothing of the edges whenever pheromone concentration levels are going below or above the extremes (see figs. 5-9). 5.3.1 eliteas fig 5. average deviation from optimum ants in relation to 101 cities (alobaedy et al., 2017) 1 % 10 % 20 % 30 % 40 % 50 60 % % 70 % 80 % 90 % 100 % 5 10 % 15 % eliteas | 101 cities 1 % 25 % 50 % 75 100 % specialists in relation to ants samuel w lusweti et al. impact of number of artificial ants in aco … | 130 fig 6. average deviation from optimum ants in relation to 225 cities (alobaedy et al., 2017) in eliteas as shown in figs. 5 and 6, only between 10-30% of the ants showed a better performance. 5.3.2. rankedas fig 7. average deviation from optimum ants in relation to 101 cities (alobaedy et al., 2017) fig 8. average deviation from optimum ants in relation to 225 cities (alobaedy et al., 2017) 1 % 10 % 20 % 30 % 40 % 50 % 60 % 70 80 % % 90 % 100 ants in relation to cities 5 % 10 % % 15 eliteas | 225 cities 1 25 % % 50 % 75 % 100 specialists in relation to ants 1 % 10 % 20 % 30 % 40 % 50 % 60 % 70 80 % % 90 % 100 5 % 10 % 15 % 1 25 % 50 % 75 % 100 % 5 specialists in relation to ants 1 % 10 20 % 30 % 40 % 50 % 60 % 70 % % 80 % 90 % 100 ants in relation to cities 5 % % 10 15 % rankedas | 225 cities 1 % 25 % 50 % 75 % 100 5 specialists in relation to ants 131 |international journal of informatics information system and computer engineering 3(1) (2022) 124-135 fig 9. average deviation from optimum ants in relation to 532 cities (alobaedy et al., 2017) as shown from the results in figs. 5,6,7,8 and 9 above, the results of rankedas and eliteas are contrasting each other. a large percentage of specialized ants degrades the solution more than a low proportion of specialized ants. in rankedas, the number of normal ants has an effect of deviation from optimum solution when using 5 specialized ants but when using more specialized ants there is no effect. in this case, the optimal solution is obtained when 5 specialized ants and 100% regular ants are used in relation to cities (alobaedy et al., 2017). nonetheless, with figure 9 where we have 525 cities, using more than 50% of the normal ants brings about worse results as the deviation from optimum is high as shown by the red line. it found that, when implementing some rankedas, the highest number of normal ants which is 100% of ants showed better performance in terms of convergence (see figs. 10-12). 5.3.3 min-max ant system fig 10. average deviation from optimum against ants in relation to 101 cities (alobaedy et al., 2017) 1 % 10 % 20 30 % 40 % % 50 % 60 % 70 80 % % 90 % 100 ants in relation to cities 5 % % 10 % 15 % 20 25 % rankedas | 532 cities 1 25 % % 50 % 75 100 % 5 specialists in relation to ants 1 % 10 % 20 30 % 40 % % 50 60 % 70 % % 80 % 90 % 100 ants in relation to cities % 2 4 % % 6 mmas | 101 cities average best-worst samuel w lusweti et al. impact of number of artificial ants in aco … | 132 fig 11. average deviation from optimum against ants in relation to 225 cities (alobaedy et al., 2017) fig 12. average deviation from optimum against ants in relation to 532 cities (alobaedy et al., 2017) fig. 10 and 11 show that increasing the number of ants in mmas has no effect on the deviation from optimum results. except for fig. 12 which shows some variation, overall performance of mmas has is not affected by a large number of ants (alobaedy et al., 2017). 6. results and discussion while some eight experiments in the data collected show that an increase in the number of ants degrades the optimization of various aco versions, two of them indicate that it actually improves the optimization of the algorithm, yet two more reveal that it has no impact on the results of the algorithm. to help understand the variation in eliteas, we single out figure 9 that brought about worse results when the number of normal ants is increased. first, we look at the number of specialized ants kept at 5 in all the figures 7, 8 and 9. calculating the ratio between the number specialized ants to that of normal ants in all the three experiments when the 1 % 10 % 20 % 30 40 % % 50 60 % 70 % 80 % % 90 % 100 ants in relation to cities 0 % % 2 % 4 6 % mmas | 225 cities average best-worst 1 % 10 % 20 % 30 40 % 50 % 60 % 70 % % 80 % 90 % 100 ants in relation to cities % 6 8 % % 10 % 12 % 14 mmas | 532 cities average best-worst 133 |international journal of informatics information system and computer engineering 3(1) (2022) 124-135 algorithm is at its optimum performance we find; 5:100, 5:225 and 5:266 (at 50% ants) for figures 7,8 and 9 respectively. when we simplify the ratio, we get 1:20, 1:45 and 1:53 respectively. if we take a look at the same figure 9, at 60% ants where the graph starts to deviate from the optimum, the ratio of the ants is 1: 63. from 60%, the graphs exponentially continue to deviate from the optimal solution till 100%. from this data we can see that for the optimal functioning of the algorithm the ratio of specialized ants and that of normal ants has to be considered. in figure 9, this ratio was not considered as the number of specialized ants remained to be 5. for instance, for every 1 specialized ant, there needs to be about 20 to about 50 normal ants to optimize the solution. when the aco algorithms are optimized, they can then be applied in various fields of study. in this case when it is applied in computer networks, the packets would be translated into ants and made to be as intelligent as the ants of the algorithm, hence prevent packet looping and many other problems associated with unoptimized computer networks especially under dynamic situations. this helps improve on the network convergence time without making the time too short to cause premature convergence or too long to bring a lot of latencies in the network. 5. conclusion after identifying the main challenges faced by aco which include stagnation of ants, actual number of routing messages that are needed, and convergence time and their main causes which are associated with parameter setting in the algorithm, we singled out one of the parameters which is the convergence time influenced by the number of ants in the solution space. however, the key problem is determining the ideal number of ants to utilize in the algorithm. it is difficult to quantify the number of ants necessary to solve a issue. first, some problems are less complex than others that means they need a smaller number of ants to be solved, and secondly, we have different types of aco algorithms working differently. a well optimized algorithm will have a short convergence time. we therefore considered the results from the experiments done by various researchers as shown in the 12 figures above and came up with the following conclusions that can help determine the optimum number of ants needed. 1) the type of aco algorithm in use, 2) the complexity of the problem under study, and 3) the ratio of the number of specialized ants to that of normal ants. further research needs to be carried out to determine how best we can calculate and determine the optimal number of ants needed. references liao, t., stützle, t., de oca, m. a. m., & dorigo, m. (2014). a unified ant colony optimization algorithm for continuous optimization. european journal of operational research, 234(3), 597-609. blum, c. (2005). ant colony optimization: introduction and recent trends. physics of life reviews, 2(4), 353-373. samuel w lusweti et al. impact of number of artificial ants in aco … | 134 kar, a. k. (2016). bio inspired computing–a review of algorithms and scope of applications. expert systems with applications, 59, 20-32. adiga, c. s., joshi, h. g., & harish, s. v. (2013). network routing problem-a simulation environment using intelligent technique. international journal of advanced computer research, 3(4), 166. dressler, f., & akan, o. b. (2010). a survey on bio-inspired networking. computer networks, 54(6), 881-900. gustov, v., & levina, a. (2021, april). electromagnetic fields as a sign of sidechannel attacks in gsm module. in 2021 11th ifip international conference on new technologies, mobility and security (ntms) (pp. 1-5). ieee. sim, k. m., & sun, w. h. (2002, november). multiple ant-colony optimization for network routing. in first international symposium on cyber worlds, 2002. proceedings. (pp. 277-281). ieee. kamali, s., & opatrny, j. (2007, march). posant: a position-based ant colony routing algorithm for mobile ad-hoc networks. in 2007 third international conference on wireless and mobile communications (icwmc'07) (pp. 21-21). ieee. kar, a. k. (2016). bio inspired computing–a review of algorithms and scope of applications. expert systems with applications, 59, 20-32. di caro, g., & dorigo, m. (2004). ant colony optimization and its application to adaptive routing in telecommunication networks (doctoral dissertation, phd thesis, faculté des sciences appliquées, université libre de bruxelles, brussels, belgium). jacobsen, r. h., zhang, q., & toftegaard, t. s. (2011). bioinspired principles for largescale networked sensor systems: an overview. sensors, 11(4), 4137-4151. caro, g. d., & dorigo, m. (1998, september). ant colonies for adaptive routing in packet-switched communications networks. in international conference on parallel problem solving from nature (pp. 673-682). springer, berlin, heidelberg. selvi, v., & umarani, r. (2010). comparative analysis of ant colony and particle swarm optimization techniques. international journal of computer applications, 5(4), 16. caro, g. d., & dorigo, m. (1998, september). ant colonies for adaptive routing in packet-switched communications networks. in international conference on parallel problem solving from nature (pp. 673-682). springer, berlin, heidelberg. wei, x. (2014). parameters analysis for basic ant colony optimization algorithm in tsp. international journal of u-and e-service, science and technology, 7(4), 159170. colorni, a., dorigo, m., & maniezzo, v. (1991). distributed optimization by ant colonies. in proceedings of ecal91–european conference on artificial life. sivagaminathan, r. k., & ramakrishnan, s. (2007). a hybrid approach for feature subset selection using neural networks and ant colony optimization. expert systems with applications, 33(1), 49-60. 135 |international journal of informatics information system and computer engineering 3(1) (2022) 124-135 yilmaz, a. h., & aydin, z. examination of parameters used in ant colony algorithm over truss optimization. balıkesir üniversitesi fen bilimleri enstitüsü dergisi, 24(1), 263-280. alobaedy, m. m., khalaf, a. a., & muraina, i. d. (2017, may). analysis of the number of ants in ant colony system algorithm. in 2017 5th international conference on information and communication technology (icoic7) (pp. 1-5). ieee. pettersson, l., & lundell johansson, c. (2018). ant colony optimization-optimal number of ants. bullnheimer, b., hartl, r. f., & strauss, c. (1997). a new rank based version of the ant system. a computational study. stützle, t., & hoos, h. h. (2000). max–min ant system. future generation computer systems, 16(8), 889-914. aydin, z. izgara si̇stemleri̇n opti̇mi̇zasyonu üzeri̇nden karinca koloni̇ opti̇mi̇zasyon algori̇tmasinda karinca sayisinin beli̇rlenmesi̇. uludağ üniversitesi mühendislik fakültesi dergisi, 22(3), 251262. and signal processing, 2(3), 1-10. ivan kristianto singgih. agricultural drone zoning and deployment strategy with multiple...| 66 agricultural drone zoning and deployment strategy with multiple flights considering takeoff point reach distance minimization ivan kristianto singgih quantum machine learning laboratory, school of industrial management engineering, korea university, seoul, republic of korea. *corresponding email: 1ivanksinggih@gmail.com a b s t r a c t s a r t i c l e i n f o in the agricultural sector, drones are used to spray chemicals for the plants. a lawn mowing movement pattern is one of the widely used methods when deploying the drones because of its simplicity. a route planner determines some pre-set routes before making the drones to fly based on them. each drone flight is limited by its battery level or level of spray liquids. to efficiently complete the spraying task, multiple drones need to be deployed simultaneously. in this study, we study a multiple drone zoning and deployment strategy that minimizes the cost to set up equipment at the takeoff points, e.g., between flights. we propose a method to set the flight starting points and directions appropriately, given various target areas to cover. this is the first study that discusses the spraying drone zoning and deployment plan while minimizing the number of takeoff points, which plays an important role in reducing the drone set up and deployment costs. the suggested procedure helps drone route planners to generate good routes within a short time. the generated routes could be used by the planner for their chemical spraying activity and could be used as initial input for their design, which can be improved with the planners’ experience. our study shows that when generating an efficient route, we must consider the number of flight area levels, directions of the drone movements, the number of u-turns of the drones, and the start points of the drone flights. article history: received 18 dec 2021 revised 20 dec 2021 accepted 25 dec 2021 available online 26 dec 2021 aug 2018 __________________ keywords: drone, routing, rule, spraying, zone control. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(2) (2021) 66-79 67 | international journal of informatics information system and computer engineering 2(2) (2021) 66-79 1. introduction drones is one of the most emerging technologies in industry 4.0 era (fernández-caramés et al., 2019). they are used in various fields, including postdisaster observation (singgih et al., 2020), last-mile delivery (moadab et al., 2022), inspection (nguyen et al., 2018), medical treatment (pulver and wei, 2018), etc. in agricultural field, drones are used for various purposes, including vegetation segmentation (su et al., 2021), soil mapping, disease monitoring, estimating crop yield (maddikunta et al., 2021), cultivation planning (kakamoukas et al., 2022), etc. in contrast to manual work, using drones allows us to perform high precision work within a short time. such great precision is enabled internet of things and big data technology (patrik et al., 2019; sordan et al., 2022). although the drone technology must still need to be supported with the establishment of internet connectivity for farmers to ensure the effective implementation of the technology (mehrabi et al., 2021), currently, there is a quick development of the drone technology itself (santangeli et al., 2020). in this study, we focus on drones that are used to spray chemicals on a large area (figure 1). before using the drones to spray chemicals on the plants, a drone route plan (figure 2) is required (hobby hangar, 2022). to assess the novelty of our study, we conducted a literature review as follows. we start with (chung et al. 2020). a drone routing problem can be classified into routing through target points and routing through target areas. in the first classification, the drones must visit all given target points (coutinho et al., 2018; gu et al., 2020), e.g., to conduct item delivery or observation from such points, while in the latter classification, the drones need to cover all target areas, e.g., when spraying chemicals on a target agriculture field (faiçal et al., 2017). our studied problem is classified as the latter one; thus, we review studies listed in table 4 (drone routing problems with area coverage classification) from (chung et al. 2020), specifically the ones marked with ag (agriculture) as the application area. to ensure that we cover papers published after (chung et al. 2020), we also searched for papers citing (chung et al. 2020). among a total of 24 papers from both sources, we found five related papers to be compared with ours. the reasons for excluding other studies are because they study target point visiting problems or focus on the non-agriculture fields (e.g., traffic monitoring and delivery system). comparisons between our study and those five related papers are presented in table 1. based on our knowledge, our study is the first one that considers drone zoning and routing when using multiple flights for agricultural purposes while minimizing the number of takeoff points. the drones’ lawn moving is a sweeping movement (otto et al., 2018; avellar et al., 2015). such movements could be differentiated into (1) movements parallel to the longest side of the area and (2) movements perpendicular to the longest side of the area, as shown in figure 3. such lawn moving is preferred when the target search area is large, in contrast to spiral movement (cabreira et al., 2019). having such a simple movement pattern allows the planner to route the drones easily while ensuring high effectiveness in the drones’ movement. our study focuses on proposing a certain simple movement to assist the route planner with their manual routing procedure. also, our proposed movement strategy could be used as a ivan kristianto singgih. agricultural drone zoning and deployment strategy with multiple...| 68 built-in route suggestion from the routing application that would be provided to the route planners for editing and approval. given such a simple yet effective movement pattern, it would be easily understood by the route planners allowing them to conduct a better route optimization. our study differs from previous studies by proposing a simple lawn moving pattern that considers the distances between each end point of a travel with the start point of the next travel (take off point). such consideration is important because the drones would need their battery and chemical container to be replaced before continuing their next travel (kim and lim, 2018; qin et a., 2021; jorge et al., 2021). minimizing the distances between take-off points minimizes the time required by the workers to pre-place the battery and chemical container replacements; thus, it significantly reduces the operational time and, in the end, reduces the cost and increases the benefit when using the drones. figure 1. chemical spraying drone source: ahmed et al. (2021) figure 2. a pre-set drone route for chemical spraying on an agricultural area source: hobby hangar (2022) table 1. comparison with previous studies characteristics drone flight routing direction(s) details on routes routing objective(s) moon and shim (2009) multiple complex exist completion time barrientos et al. (2011) multiple complex exist completion time avellar et al. (2015) multiple lawn mowing (1 direction) exist completion time barna et al. (2019) single lawn mowing (1 direction) exist captured photo quality 69 | international journal of informatics information system and computer engineering 2(2) (2021) 66-79 tu et al. (2020) multiple lawn mowing (1 direction), grid not exist captured photo quality our study multiple lawn mowing (2 directions) exist completion time, number of takeoff points (a) parallel to the longest side (b) perpendicular to the longest side figure 3. two types of lawn moving patterns of the drones when solving routing problems, various solution approaches could be used, e.g., mathematical models, exact heuristics, metaheuristics, simulation, and rule-based methods (amarat and zong, 2019; erdelić and carić, 2019; moghdani et al., 2021). lately, machine learning-based methods are also proposed by arnold and (sörensen, 2019) and zhao, et al. 2021). earlier methods (e.g., exact methods) produce better solution quality but require more computational effort. in contrast, rulebased methods are straightforward and can be applied more easily. development of such rule-based methods is common for solving various combinatorial optimization problems, e.g., project scheduling (chakrabortty et al., 2020), job dispatching (ðurasević and jakobović, 2020), and machine scheduling (gil-gala et al., 2019). related to our problem, we develop a rule-based approach to provide drone route planners with the necessary insights for manually designing the routes. the structure of the whole paper is presented as follows: section 2 explains the proposed routing procedure. section 3 presents the numerical experiments and discussions. finally, section 4 concludes the study. 2. method when operating the drones, we need to ensure real drone characteristics, e.g., the limitations on flight range (otto et al., 2018) and limitations on weight to carry (macrina et al., 2020). to ensure each drone to completes its tasks, we need to set up some locations within the working area of the drone with the equipment necessary to conduct the battery charging and chemical refuelling or replacement. when any drone requires a temporary landing, the required equipment must be ivan kristianto singgih. agricultural drone zoning and deployment strategy with multiple...| 70 ready. please note that after the landing, the drones would take off again after the landing to continue the spraying process. therefore, we call the temporary landing points as take-off points as well. to ensure the readiness of the equipment, it is straightforward to minimize the distances between the take-off points. such a distance minimization can also be found in truck routing problems in a truck-drone collaborative parcel delivery system (wang et al., 2019). such a take-off point distance minimization is equivalent to reducing the number of take-off points, which reduces the effort to transport and prepare the equipment for the landing drones. our proposed drone zoning and deployment procedure is described in algorithm 1. algorithm 1. drone zoning and deployment procedure 1: calculate the number of required drone flights: " #_of_drone_flights=" ⌊"total grid area/max covered grid area per drone" ⌋ 2: determine alternatives of identical drone flight area dimensions (length and width), which cover the whole spraying area 3: define the possible drone flight start and end positions simultaneously while minimizing the total distances from the drone take-off points 4: finalize the best drone deployment plan the end points for each drone movement are determined based on the size of the target area and the movement direction of the drones, as shown in figure 3. as shown in figure 3(a), when the number of u-turns of a drone is even, the drone travel will end at on the exact opposite side of the starting point. meanwhile, when the number of u-turns of the drone is odd (figure 3(b)), the drone travel will end at on the same side with as the starting point. considering such a movement rule, we need to determine the movement direction of the drones based on the size of the target area. it will significantly affect the positions of the take-off points and determine the number of the take-off points. in general, minimizing the number of u-turns is preferred because travelling through the u-turn area causes a longer movement time due to the required deceleration and acceleration movements. however, allowing a decent number of u-turns should be acceptable, considering that making such decisions could reduce the number of take-off points. please refer to the next section for examples and further analysis. 3. result and disscusion for the numerical experiment, we consider two problem instances. instance 1 considers a 200-grid area with 10 grid x 20 grid dimensions, while instance 2 considers a 180-grid area with 15 grid x 12 grid dimensions. for instances 1 and 2, the max grid area covered by a drone flight are 20 and 30 grid area, respectively. considering various movement rules, we generate five and three drone deployment plans for instances 1 and 2, as shown in figures 4 and 5, respectively. we observe the drone flight area/drone movements on 71 | international journal of informatics information system and computer engineering 2(2) (2021) 66-79 horizontal and vertical locations/directions for simplicity. please note that we refer to the horizontal or vertical directions when observing the target area from the top view. the results are produced using algorithm 1. in step 1, it is straightforward to determine the number of flights based on the drone’s limited flight time, which is determined by the limited battery or carried chemical. in step 2, we define the same-shaped flight area for the drones. currently, we consider the same shape to extract basic drone deployment and routing rules easily. to observe various routing alternatives in steps 3 and 4, we test various drone movement strategies, e.g., (1) horizontal or vertical-directed movements and (2) starting points at the outer side or inner side of the target area. for a drone flight, the start point, and end point are labeled as “s” and “e”, respectively. a “takeoff point” consists of a maximum of 4 landing and takeoff points that are placed adjacently because we can place the equipment (recharged battery and refuel chemical tanks) in the center of those points. from this part of the manuscript, we will call each location as “a takeoff point”. we exclude the first start point and the last end point from the calculation for the number of takeoff points because we assume that each drone is ready with all required equipment when starting its first flight and at the end of its last flight, we do not need to rush with the drone last pickup process. as an example, in figure 4(a), there are five takeoff points as follows: takeoff point 1 (e1, s2, e9, s10), takeoff point 2 (e2, s3, e8, s9), takeoff point 3 (e3, s4, e7, s8), takeoff point 4 (e4, s5, e6, s7), and takeoff point 5 (e5, s6). different background colors for each takeoff point group are used to clearly present the results. (a) even horizontal flight area, vertical flight, odd u-turns, 5 takeoff points (best) ivan kristianto singgih. agricultural drone zoning and deployment strategy with multiple...| 72 (b) even horizontal flight area, vertical flight, odd u-turns, 9 takeoff points (c) even horizontal flight area, horizontal flight, even u-turns, 7 takeoff points (d) odd horizontal flight area, horizontal flight, odd u-turns, 5 takeoff points 73 | international journal of informatics information system and computer engineering 2(2) (2021) 66-79 (e) even horizontal flight area, horizontal flight, no u-turns, 9 takeoff points figure 4. drone deployment alternatives for instances 1 (a) odd horizontal flight area, vertical flight, odd u-turns, 5 takeoff points ivan kristianto singgih. agricultural drone zoning and deployment strategy with multiple...| 74 (b) odd horizontal flight area, vertical flight, odd u-turns, 5 takeoff points (c) odd horizontal flight area, horizontal flight, even u-turns, 4 takeoff points (best) figure 5. drone deployment alternatives for instances 2 the best routing alternatives (that reduces the number of take-off points) for instances 1 and 2 are shown in figures 4(a) and 5(c), respectively. based on our observation, we conclude that a minimum number of take-off points can be produced by: (1) generating an even number of flight area levels, then ensuring that the take-off points are grouped in the 75 | international journal of informatics information system and computer engineering 2(2) (2021) 66-79 middle of each adjacent two levels, as shown by figure 4(a), which has two horizontal flight area levels. the takeoff points can be accumulated in the middle of both horizontal levels because the number of u-turns is odd (which makes the drones return to the same middle side of the target area). (2) generating an odd number of flight area levels, then setting the drones’ perpendicular movements from those levels, as shown by figure 5(c), which has three horizontal flight area levels and a horizontal drone flight movement. in this example, the number of u-turns is even. if the number of u-turns is odd, then we would follow the same routing solution structure shown in figure 4(a) by starting the drone movements from the middle part of two adjacent vertical flight area. in addition to the findings above, we also conclude that no u-turns do not minimize the number of take-off points because the start and end points are not adjacently placed. the findings in this study can be used as a good reference for drone flight planners when they predefine the flight routes (ma et al., 2019). 4. conclusion we study a multiple drone zoning and deployment problem. our proposed method includes dividing the target area into zones, then determining detailed drone movement directions to minimize the effort of preparing battery and refueling chemicals for drones at the end of each flight. some useful insights are listed to be used as recommendations for drone flight planners. for future research topics, we suggest allowing different drone flight area sizes to increase the flexibility of the drone movements and the effort for the equipment preparation. references ahmed, s., qiu, b., ahmad, f., kong, c.-w., & xin, h. (2021). a state-of-the-art analysis of obstacle avoidance methods from the perspective of an agricultural sprayer uav’s operation scenario. agronomy, 11, 1069. amarat, s. b., & zong, p. (2019). 3d path planning, routing algorithms and routing protocols for unmanned air vehicles: a review. aircraft engineering and aerospace technology, 91(9), 1245–1255. arnold, f., & sörensen, k. (2019). what makes a vrp solution good? the generation of problem-specific knowledge for heuristics. computers & operations research, 106, 280–288. avellar, g. s. c., pereira, g. a. s., pimenta, l. c. a., & iscold, p. (2015). multi-uav routing for area coverage and remote sensing with minimum time. sensors, 15, 27783–27803. ivan kristianto singgih. agricultural drone zoning and deployment strategy with multiple...| 76 barna, r., solymosi, k., & stettner, e. (2019). mathematical analysis of drone flight path. journal of agricultural informatics, 10(2), 15–27. barrientos, a., colorado, j., del cerro, j., martinez, a., rossi, c., sanz, d., & valente, j. (2011). aerial remote sensing in agriculture: a practical approach to area coverage and path planning for fleetsof mini aerial robots. journal of field robotics, 28(5), 667–689. cabreira, t. m., brisolara, l. b., & paulo r. jr., f. (2019). survey on coverage path planning with unmanned aerial vehicles. drones, 3(1), 4. chakrabortty, r. k., rahman, h. f., & ryan, m. j. (2020). efficient priority rules for project scheduling under dynamic environments: a heuristic approach. computers & industrial engineering, 140, 106287. chung, s. h., sah, b., & lee, j. (2020). optimization for drone and drone-truck combined operations: a review of the state of the art and future directions. computers and operations research, 123, 105004. coutinho, w. p., battara, m., & fliege, j. (2018). the unmanned aerial vehicle routing and trajectory optimisation problem, a taxonomic review. computers & industrial engineering, 120, 116–128. ðurasević, m., & jakobović, d. (2020). comparison of schedule generation schemes for designing dispatching rules with genetic programming in the unrelated machines environment. applied soft computing journal, 96, 106637. erdelić, t., & carić, t. (2019). a survey on the electric vehicle routing problem: variants and solution approaches. journal of advanced transportation, 5075671. faiçal, b. s., freitas, h., gomes, p. h., mano, l. y., pessin, g., de carvalho, a. c. p. l. f., krishnamachari, b., & ueyama, j. (2017). an adaptive approach for uavbased pesticide spraying in dynamic environments. computers and electronics in agriculture, 138, 210–223. fernández-caramés, t. m., blanco-novoa, o., froiz-míguez, i., & fraga-lamas, p. (2019). towards an autonomous industry 4.0 warehouse: a uav and blockchain-based system for inventory and traceability applications in big data-driven supply chain management. sensors, 19(10), 2394. gil-gala, f. j., mencía, c., sierra, m. r., & varela, r. (2019). evolving priority rules for on-line scheduling of jobs on a single machine with variable capacity over time. applied soft computing journal, 85, 105782. 77 | international journal of informatics information system and computer engineering 2(2) (2021) 66-79 gu, q., fan, t., pan, f., & zhang, c. (2020). a vehicle-uav operation scheme for instant delivery. computers & industrial engineering, 149, 106809. hobby hangar. (2022). align rm41501xw m4 high-performance agricultural drone. https://www.hobbyhangar.co.nz/align-rm41501xw-m4-highperformance-agricultural-drone. jorge, h. g., de santos, l. m. g., álvarez, n. f., sánchez, j. m., & medina, f. n. (2021). operational study of drone spraying application for the disinfection of surfaces against the covid-19 pandemic. drones, 5(1), 18. kakamoukas, g. a., sarigiannidis, p. g., & economides, a. a. (2022). fanets in agriculture a routing protocol survey. internet of things, 18, 100183. kim, s. j., & lim, g. j. (2018). a hybrid battery charging approach for drone-aided border surveillance scheduling. drones, 2(4), 38. ma, f., xu, z., & xiong, f. (2019). research on route planning of plant protection uav based on area modular division. 2019 11th international conference on intelligent human-machine systems and cybernetics (ihmsc) (p. 101–104). macrina, g., pugliese, l. d. p., guerriero, f., & laporte, g. (2020). drone-aided routing: a literature review. transportation research part c, 120, 102762. maddikunta, p. k. r., hakak, s., alazab, m., bhattacharya, s., gadekallu, t. r., khan, w. z., & pham, q.-v. (2021). unmanned aerial vehicles in smart agriculture: applications, requirements, and challenges. ieee sensors journal, 21(16), 17608–17619. mehrabi, z., mcdowell, m. j., ricciardi, v., levers, c., martinez, j. d., mehrabi, n., wittman, h., ramankutty, n., & jarvis, a. (2021). the global divide in datadriven farming. nature sustainability, 4, 154–160. moadab, a., farajzadeh, f., & valilai, o. f. (2022). drone routing problem model for last-mile delivery using the public transportation capacity as moving charging stations. scientific reports, 12, 6361. moghdani, r., salimifard, k., demir, e., & benyettou, a. (2021). the green vehicle routing problem: a systematic literature review. journal of cleaner production, 279, 123691. moon, s.-w., & shim, h.-c. (2009). study on path planning algorithms for unmanned agricultural helicopters in complex environment. international journal of aeronautical & space sciences, 10(2), 1–11. ivan kristianto singgih. agricultural drone zoning and deployment strategy with multiple...| 78 nguyen, v. n., jenssen, r., & roverso, d. (2018). automatic autonomous visionbased power line inspection: a review of current status and the potential role of deep learning. international journal of electrical power & energy systems, 99, 107–120. otto, a., agatz, n., campbell, j., golden, b., & pesch, e. (2018). optimization approaches for civil applications of unmanned aerial vehicles (uavs) or aerial drones: a survey. networks, 72(4), 411–458. patrik, a., utama, g., gunawan, a. a. s., chowanda, a., suroso, j. s., shofiyanto, r., & budiharto, w. (2019). gnss-based navigation systems of autonomous drone for delivering items. journal of big data, 6, 53. pergher, i., frej, e. a., roselli, l. r. p., & de almeida, a. t. (2020). integrating simulation and fitradeoff method for scheduling rules selection in job-shop production systems. international journal of production economics, 227, 107669. pulver, a., & wei, r. (2018). optimizing the spatial location of medical drones. applied geography, 90, 9–16. qin, y., kishk, m. a., & alouini, m.-s. (2021). on the influence of charging stations spatial distribution on aerial wireless networks. ieee transactions on green communications and networking, 5(3), 1395–1409. santangeli, a., chen, y., kluen, e., chirumamilla, r., tiainen, j., & loehr, j. (2020). integrating drone‑borne thermal imaging with artifcial intelligence to locate bird nests on agricultural land. scientific reports, 10, 10993. singgih, i. k., lee, j., & kim, b.-i. (2020). node and edge drone surveillance problem with consideration of required observation quality and battery replacement. ieee access, 8, 44125–44139. sordan, j. e., oprime, p., pimenta, m. l., chiabert, p., & lombardi, f. (2022). industry 4.0: a bibliometric analysis in the perspective of operations management. operations and supply chain management, 15(1), 93–104. su, j., yi, d., su, b., mi, z., liu, c., hu, x., xu, x., guo, l., & chen, w.-h. (2021). aerial visual perception in smart farming: field study of wheat yellow rust monitoring. ieee transactions on industrial informatics, 17(3), 2242–2249. tu, y.-h., phinn, s., johansen, k., robson, a., & wu, d. (2020). optimising drone flight planning for measuring horticultural tree crop structure. isprs journal of photogrammetry and remote sensing, 160, 83–96. 79 | international journal of informatics information system and computer engineering 2(2) (2021) 66-79 wang, d., hu, p., du, j., zhou, p., deng, t., & hu, m. (2019). routing and scheduling for hybrid truck-drone collaborative parcel delivery with independent and truck-carried drones. ieee internet of things journal, 6(6), 10483–10495. zhao, j., mao, m., zhao, x., & zou, j. (2021). a hybrid of deep reinforcement learning and local search for the vehicle routing problems. ieee transactions on intelligent transportation systems, 22(11), 7208–7218. 21 | international journal of informatics information system and computer engineering 3(2) (2022) 21-30 2-d attention-based convolutional recurrent neural network for speech emotion recognition akalya devi c, karthika renuka d, aarshana e winy, p c kruthikkha, ramya p, soundarya s assistant professor, 2ug scholar, department of information technology,psg college of technology, coimbatore, india *corresponding email: cad.it@psgtech.ac.in a b s t r a c t s a r t i c l e i n f o recognizing speech emotions is a formidable challenge due to the complexity of emotions. the function of speech emotion recognition (ser) is significantly impacted by the effects of emotional signals retrieved from speech. the majority of emotional traits, on the other hand, are sensitive to emotionally neutral elements like the speaker, speaking manner, and gender. in this work, we postulate that computing deltas for individual features maintain useful information which is mainly relevant to emotional traits while it minimizes the loss of emotionally irrelevant components, thus leading to fewer misclassifications. additionally, speech emotion recognition (ser) commonly experiences silent and emotionally unrelated frames. the proposed technique is quite good at picking up important feature representations for emotion relevant features. so here is a two dimensional convolutional recurrent neural network that is attention-based to learn distinguishing characteristics and predict the emotions. the mel-spectrogram is used for feature extraction. the suggested technique is conducted on iemocap dataset and it has better performance, with 68% accuracy value. article history: received 18 dec 2022 revised 20 dec 2022 accepted 25 dec 2022 available online 26 dec 2022 aug 2018 __________________ keywords: 2-d, attention-based, convolutional recurrent neural network, speech emotion recognition international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 21-30 akalya devi et al. 2-d attention-based convolutional recurrent neural network … | 22 1. introducrion the significance of human speech emotion recognition has increased recently to increase the quality and efficiency of interactions between machines and humans (khalil et al., 2019). due to the difficulty in defining both natural and artificial emotions, recognizing human emotions is a challenging task all on its own. extraction of the spectral and prosodic elements that would lead to the accurate assessment of emotions has been the subject of numerous investigations (tzirakis et al., 2018). recognition of speech emotions is a technique that uses a processor to extract emotional information from speech signals (chen et al., 2018). it then compares and analyzes the collected emotional information together with the distinctive factors. after the emotional information is extracted, various techniques and concepts are used to predict the emotions of speech signals (khalil et al., 2019). speech emotion detection is now a rapidly developing discipline bridging the interaction between robots and humans. it is also a popular study area in signal processing and pattern recognition. emotions are incredibly important to human mental health. it is a way of expressing one's thoughts or state of mind to others. the major objective of ser is to improve human-machine interaction (hmi). it can also be used with lie detectors to monitor a subject's psychophysical state (lalitha et al., 2015). onboard car driving systems, dialogue systems for spoken languages used in call center conversations and the utilization of speech emotion patterns in medical applications are a few instances of ser. hmi systems still have a lot of problems that need to be resolved, especially when they are shifted from being tested in labs to being used in actual operations. therefore, efforts are required to effectively resolve all these problems and enhance machine emotion perception. recently, deep neural networks (dnns) have gained popularity and made revolutionary strides in a number of machines learning fields, including the continuous effect identification field. in most of the studies, hand-crafted features are used to feed the dnn architectures. many dnn architectures have been put forth in that approach, including convolutional neural networks (cnns) and long-short term memory (lstm) networks. mao et al. (mao et al., 2014) first used convolutional neural networks (cnn) and demonstrated great scores on numerous benchmark datasets for learning affective-salient features for ser. recurrent neural networks (rnns) were used by lee et al. (lee & tashey, 2015) to train ser on long-range temporal correlations. in order to train a convolutional recurrent neural network (crnn) to predict continuous valence space, trigeorgis et al. (trigeorgis et al., 2016) directly used the raw audio data. additionally, structures connecting the output and the input segments have been learned with significant effectiveness using attention mechanismbased rnns. rnns based on attention mechanisms are ideally suited to the ser tasks. first, speech is basically a sequence of data with different lengths. the majority of speech signals annotate emotion labels at the utterance level even though utterances sometimes have lengthy pauses and frequently have a 23 | international journal of informatics information system and computer engineering 3(2) (2022) 21-30 short word count. selecting emotionalrelevant frames for ser is very crucial. in this paper, we extend our model to yield affective salient characteristics for the final emotion categorization using a crnn-based attention mechanism. in this study, we combine crnn and an attention model to create a unique architecture for ser dubbed 2-d attention-based convolutional recurrent neural networks (acrnn). the following is a summary of this paper's main contributions: 1) we suggest a unique 2-d crnn for ser that enhances the ability to understand the time-frequency relationship. 2) we employ an additional attention model to automatically concentrate on the emotion-crucial frames and offer discriminative utterance-level characteristics for ser to cope with silence critical frames and emotionirrelevant frames. 3) experimental results contain athe ccuracy, recall, precision and confusion matrix of our proposed model. it is well known that most speech emotion datasets only have utterancelevel class labels. most sentences, however, contain silent regions, short pauses, transitions between phonemes, unvoiced phonemes, and so on. it is clear that not all parts of a sentence are emotionally connected. unfortunately, lstm does not handle this situation well when analyzing acoustic characteristics extracted from voice. in the current study, emotion classification is useful for distinguishing between emotionallyrelevant and emotionally-irrelevant frames. in emotion classification, it is useful to know whether the speech frame is voiced or unvoiced. currently, there are two types of commonly used methods: manually extracting emotionally relevant speech frames and using models to learn how to distinguish automatically. however, as manual extraction requires different thresholds on different data sets, it has some limitations in terms of feasibility. human emotional expression is often gradual and thus each voiced frame is useful for emotion classification. attention mechanisms can better match human emotional expression by capturing only the affective frames. local attention was added to lstm and different weights were assigned to each frame with varying emotional intensity. 2. literature survey on the iemocap dataset, sarthak tripathi and homayoon beigi performed multimodal emotion detection and determined the best individual architectures for classification of each modality using data from speech, text and motion capture. the design of their merged model is modular. this makes it possible to upgrade any individual model without affecting the other modalities. they utilized motion captured data and 2d convolutions in place of video recordings and 3d convolutions (tripathi et al., 2018). for the arabic dataset ksuemotions, mohammed zakariah and yaser mohammad seddiq performed speech emotion recognition. the feature extraction method made use of the timefrequency data from the spectrogram, as well as numerous modification and akalya devi et al. 2-d attention-based convolutional recurrent neural network … | 24 filtering techniques were used. although the system was tested at the file and segment levels, it was trained at the segment level (maji & swain, 2022). to automatically extract affective salient features from raw spectral data, yawei mu and luis a. hernandez gomez presented a distributed convolution neural network (cnn). from the cnn output, they then applied a bidirectional recurrent neural network (brnn) to obtain temporal information. finally, they used the attention mechanism to target the emotion-relevant portions of utterance in the brnn output sequence (jiang et al., 2021). a convolutional-recurrent neural network with multiple attention mechanisms (crnn-ma) was proposed by p. jiang, x. xu, h. tao, l. zhao, and c. zou for ser. it uses extracted melspectrums and frame-level features in parallel convolutional neural network (cnn) and long short-term memory (lstm) modules, respectively. a multidimensional attention layer and multiple self-attention layers in the cnn module on frame-level weight components (yadav et al, 2021) are some of the strategies they established for the suggested crnn-ma. yadav, o. p., bastola, l. p., and sharma, j. presented the convolutional recurrent neural network (crnn), which combines convolutional neural network (cnn) and bidirectional long shortterm memory (bilstm), to learn emotional features from log-mel scaled spectrograms of spoken utterances. convolution kernels of cnn are used to learn local features and a layer of bilstm is chosen to learn the temporal dependencies from the learnt local features. speech utterances are preprocessed to cut out distracting sounds and unnecessary information. additionally, methods for increasing the number of data samples are researched, and the best methods are chosen to improve the model's recognition rate (lim et al., 2016). without employing any conventional hand-crafted features, wootaek lim, daeyoung jang, and taejin lee developed a ser approach based on concatenated cnns and rnns. particularly for computer vision tasks, convolutional neural networks (cnns) have exceptional recognition ability. recurrent neural networks (rnns) also perform sequential data processing tasks to a great extent with high degree of success. the classification result was proven to have higher accuracy than that attained using traditional classification methods by utilizing the proposed methods on an emotional speech database (gayathri et a., 2020). silent frames and inappropriate emotional frames are frequent problems for speech emotion recognition (ser). meanwhile, the attention process has proved to be exceptionally effective at learning relevant feature representations for particular activities. using the melspectrogram with deltas and delta-deltas as input, gayathri, p., priya, p. g., sravani, l., johnson, s., and sampath, v. presented a convolutional recurrent neural networks (acrnn) based on attention to learn discriminative features for ser. finally, test results 25 | international journal of informatics information system and computer engineering 3(2) (2022) 21-30 demonstrated the viability of the suggested approach and achieved cutting-edge performance in terms of unweighted average recall (gayathri et a., 2020). 2.1. proposed models and experimental setup a convolutional recurrent neural network with a 2d attention base, serves as the proposed model for speech emotion recognition. 2.2. speech emotion recognition this section explains the proposed 2d attention based convolutional recurrent neural network. convolutional neural network, or cnn or convnet, is particularly adept at processing input with a grid-like architecture, like an image. a binary representation of visual data is a digital image. recurrent neural networks (rnns) are a type of neural network in which the results of one step are fed into the next step's computations. it employs sequential data or time series data. the convolutional recurrent neural network (crnn) model uses the outputs and hidden states of the recurrent units in each frame to extract features from the successive windows by feeding each window frame by frame into the recurrent layer. here we combine an attention mechanism together with cnn and rnn that enables easier and higherquality learning by concentrating on certain portions of the input sequence inorder to predict a particular portion of the output sequence. feature extraction is a process that converts raw data into manageable numerical features while preserving the original data's information. feature extraction when compared to using machine learning or deep learning models on the raw data directly, produces better outcomes. for the feature extraction log mel-spectrogram is used. the acrnn architecture, which combines crnn with an attention model, is used. then, as depicted in fig. 1, a fully linked layer and a softmax layer for ser are introduced. fig. 1. acrnn architecture cnn has recently demonstrated remarkable accomplishments in the ser field. the time domain and frequency domain are equally important and 2dimensional convolution performs better with less data than 1-dimensional convolution. the ser findings, however, vary greatly between speakers because of huge variation in tone, voice and other unique characteristics. the log-mels with deltas and delta-deltas act as the acrnn input to handle this variation, where the deltas and delta-deltas describe the emotional transformation process. the mel scale has a range of pitches that to the human ear, appear to be equally distant from one another. the distance in hertz between mel scale values, often known as "mels," increases as the frequency increases. mel, which stands for melody, denotes that the scale is founded on pitch comparisons. akalya devi et al. 2-d attention-based convolutional recurrent neural network … | 26 extensive tests have shown us that the mel spectrum is better compatible with the human auditory sense characteristic, which exhibits the linear distribution under 1000 hz and the logarithm growth above 1000 hz and hence this point is used to obtain the log-mel spectrum static. the link between the frequency and the mel spectrum is interrelated. a mel spectrogram renders frequencies over a specific threshold logarithmically (the corner frequency). for instance, in the spectrogram with a linear scale, the vertical space between 1,000 and 2,000 hz is half that between 2,000 and 4,000 hz. the distance between the ranges is almost the same in the mel spectrogram. similar to how we hear, this scaling makes similar low frequency sounds simpler to identify from similar high frequency noises. a frequency-domain value is multiplied by a filter bank to create the output of a mel spectrogram. when a speech signal is given with zero mean and unit variance, it is used to minimize the differences between speakers. the signal is then divided into small frames using hamming windows with a shift of 10 ms and a 25 ms duration. the power spectrum is then placed through the mel-filter bank i to produce output pi, and the output is then used to calculate the power spectrum for each frame using the discrete fourier transform (dft) i. the logarithm of pi is then used to produce the log-mels mi, as shown by (1). to determine the log-mels' deltas features, we use the following formula (2). n is often selected as (2). similarly, the delta-deltas features are calculated using the time derivative of the deltas, as seen in (3). generate a 3-d feature representation for the cnn input x σ 〖〖 r〗^(t×f×c)〗 ^ by t stands for the time (frame) length, f for the number of mel-filter banks, and c for the number of feature channels when computing the log-mels with deltas and delta-deltas. as in speech recognition [17], we set f in this task to 40 and c to 3, which stand for static, deltas, and deltadeltas, respectively. 𝑚𝑖 = 𝑙𝑜𝑔(𝑝𝑖 ) ……………………(1) 𝑚𝑖 𝑑 = 𝛴 𝑛=1 𝑁 𝑛(𝑚𝑖+𝑛 − 𝑚𝑖−𝑛 ) 2𝛴 𝑛=1 𝑁 𝑛2 ………..(2) 𝑚𝑖 𝑑𝑑 = 𝛴 𝑛=1 𝑁 𝑛(𝑚𝑖+𝑛 𝑑 − 𝑚𝑖−𝑛 𝑑 ) 2𝛴 𝑛=1 𝑁 𝑛2 ………..(3) 2.3. acrnn architecture: in this part, we integrate crnn with an attention model along with 2-d log-mels. 2-d cnn is used to perform convolution in a patch that only contains a few frames on the entire log-mels. the long shortterm memory (lstm) is then fed with 2d cnn sequential characteristics for temporal summarization. a series of high-level features are then entered into the attention layer, which outputs utterance-level features. finally, utterance-level characteristics are used as the fully connected layer input to obtain higher level features for ser. 1)crnn model: high-level features for ser are retrieved using crnn from given 2-d log-mels. the crnn used here 27 | international journal of informatics information system and computer engineering 3(2) (2022) 21-30 consists of several 2-d convolution layers, one 2-d max-pooling layer, one linear layer, and one lstm layer. each convolutional layer has a 5 x 2 filter size, with the first convolutional layer having 128 feature maps and the subsequent convolutional layers having 256 feature maps. after the first convolutional layer, we only use one max pooling layer and the pooling size is 2 x 2. the model parameters can be effectively reduced without compromising accuracy by adding a linear layer before feeding 2-d cnn features into the lstm layer. as a result, we find that the linear layer with 768 output units is appropriate when added as a dimension-reduction layer after the 2-d cnn. we perform a 2-d cnn and then feed the 2-d cnn sequence features via a bidirectional rnn with 128 cells in each direction for temporal summarization. as a result, a sequence of 256-dimensional high-level feature representations are obtained. 2)attention layer: due to the fact that not all frame-level crnn features equally contribute to the representation of speech emotion, an attention layer is employed to focus on emotion-relevant sections and produce discriminative utterance-level representations for ser. instead of only using a mean/max pooling across time, the significance of a number of high-level representations to the utterance-level emotion representations is rated using an attention model. in particular, first determine the normalized weight using a softmax function and the lstm output ht at time step t. then, as illustrated, perform a weighted sum on ht using the weights to determine the utterance-level representations (5). finally, feed the utterance-level representations through a fully connected layer with 64 output units to obtain higher level representations that help the softmax classifier map the utterance representations into n different spaces, where n is the number of emotion classes. the fully connected layer is subjected to batch normalization (gayathri et a., 2020) to expedite training and enhance generalization performance. 𝑎𝑡 = 𝑒𝑥𝑝(𝑊.ℎ𝑡) 𝛴 𝑇=1 𝑇 𝑒𝑥𝑝(𝑊.ℎ𝑡) ………(4) 𝑐 = 𝛴 𝑡=1 𝑇 𝑎𝑡 ℎ𝑡 ………….(5) we conduct ser experiments using the interactive emotional dyadic motion capture database (iemocap) to assess performance of our proposed model. there are five sessions of iemocap each having utterances having duration on an average lasting for 4.5 seconds and the rate of each sample being 16 kilohertz. every session here is presented by two speakers (a male and female) in both scripted scenes and improvised scenes. only four emotions are considered here angry, sad, happy and neutral. cross validation used here for evaluation is 10fold. out of the total ten speakers, eight speakers are chosen for training the model, one speaker is chosen for testing and the other speaker is chosen for validation. consequently, we perform each evaluation multiple times using various random seeds in order to obtain more reliable findings. we divide the signal into 3 segments which are all equal in akalya devi et al. 2-d attention-based convolutional recurrent neural network … | 28 length for improved acceleration which is parallel. we have also padded with zeros for speech utterances which are lasting less than 3 s. training set’s standard deviation and mean(global) are used for normalization of log-mels of testing and training data, with 25 ms as the size of the window and a shift of 10 ms. tensorflow and keras libraries are installed for implementation (see figs. 2-4). fig. 2. workflow for azure machine learning fig. 3. workflow for azure machine learning fig 4. classification report of 2d attention based crnn fig. 1 represents the classification report of 1d cnn lstm which has an accuracy of 56%, precision value of 59% and recall value of 56%. fig. 2 represents the classification report of temporal 2-d cnn which has an accuracy of 58%, precision value of 59% and recall value of 58%. fig. 3 represents the classification report of 2-d acrnn which has an accuracy of 68%, precision value of 67% and recall value of 68%. thus, our acrnn model’s performance is superior while compared with other models (see fig. 5). fig 5. classification report of 2d attention based crnn fig. 4 displays the confusion matrix of the acrnn model. there are four emotions 0 represents angry, 1 represents sad, 2 represents happy and 3 represents neutral. the diagonal values represent the correctly predicted values. the accuracy of our proposed model 2-d crnn is 68% which is higher than the accuracy of 1d cnn lstm and t-2d cnn. weighted precision of our model is 0.67, weighted recall of our model is 0.68. weighted f1 score of our model is 0.67. all these values are higher than the corresponding values in 1d cnn lstm and t-2d cnn. thus, our model outperforms similar ser models with greater values for all metrics. 3-d 29 | international journal of informatics information system and computer engineering 3(2) (2022) 21-30 attention based crnn implemented in the paper chen (chen et al., 2018) has average recall value of 64.74%. our 2-d attention based crnn has outperformed it with a recall value of 68% (see fig. 6). fig 6. comparison of models and their evaluation metrics fig. 5 shows the plot between models and their evaluation metrics. our model comes out to be the best in all metrics while comparing with the other two models. references chen, m., he, x., yang, j., & zhang, h. (2018). 3-d convolutional recurrent neural networks with attention model for speech emotion recognition. ieee signal processing letters, 25(10), 1440-1444. gayathri, p., priya, p. g., sravani, l., johnson, s., & sampath, v. (2020). convolutional recurrent neural networks based speech emotion recognition. journal of computational and theoretical nanoscience, 17(8), 3786-3789. huang, c. w., & narayanan, s. s. (2016, september). attention assisted discovery of sub-utterance structure in speech emotion recognition. in interspeech (pp. 1387-1391). huang, c., gong, w., fu, w., & feng, d. (2014). a research of speech emotion recognition based on deep belief network and svm. mathematical problems in engineering, 2014. jiang, p., xu, x., tao, h., zhao, l., & zou, c. (2021). convolutional-recurrent neural networks with multiple attention mechanisms for speech emotion recognition. ieee transactions on cognitive and developmental systems. khalil, r. a., jones, e., babar, m. i., jan, t., zafar, m. h., & alhussain, t. (2019). speech emotion recognition using deep learning techniques: a review. ieee access, 7, 117327-117345. akalya devi et al. 2-d attention-based convolutional recurrent neural network … | 30 lalitha, s., mudupu, a., nandyala, b. v., & munagala, r. (2015, december). speech emotion recognition using dwt. in 2015 ieee international conference on computational intelligence and computing research (iccic) (pp. 1-4). ieee. lee, j., & tashev, i. (2015, september). high-level feature representation using recurrent neural network for speech emotion recognition. in interspeech 2015. lim, w., jang, d., & lee, t. (2016, december). speech emotion recognition using convolutional and recurrent neural networks. in 2016 asia-pacific signal and information processing association annual summit and conference (apsipa) (pp. 1-4). ieee. maji, b., & swain, m. (2022). advanced fusion-based speech emotion recognition system using a dual-attention mechanism with conv-caps and bi-gru features. electronics, 11(9), 1328. mao, q., dong, m., huang, z., & zhan, y. (2014). learning salient features for speech emotion recognition using convolutional neural networks. ieee transactions on multimedia, 16(8), 2203-2213. nwe, t. l., foo, s. w., & de silva, l. c. (2003, december). detection of stress and emotion in speech using traditional and fft based log energy features. in fourth international conference on information, communications and signal processing, 2003 and the fourth pacific rim conference on multimedia. proceedings of the 2003 joint (vol. 3, pp. 1619-1623). ieee. trigeorgis, g., ringeval, f., brueckner, r., marchi, e., nicolaou, m. a., schuller, b., & zafeiriou, s. (2016, march). adieu features? end-to-end speech emotion recognition using a deep convolutional recurrent network. in 2016 ieee international conference on acoustics, speech and signal processing (icassp) (pp. 5200-5204). ieee. tripathi, s., tripathi, s., & beigi, h. (2018). multi-modal emotion recognition on iemocap dataset using deep learning. arxiv preprint arxiv:1804.05788. tzirakis, p., zhang, j., & schuller, b. w. (2018, april). end-to-end speech emotion recognition using deep neural networks. in 2018 ieee international conference on acoustics, speech and signal processing (icassp) (pp. 5089-5093). ieee. yadav, o. p., bastola, l. p., & sharma, j. (2021). speech emotion recognition using convolutional recurrent neural network. 31 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 a computational bibliometric analysis of esport management using vosviewer selvia lorena br ginting faculty of information science and technology, universiti kebangsaan malaysia, malaysia corresponding email: selvialorena@yahoo.com a b s t r a c t s a r t i c l e i n f o the project aims to combine visualization study with vosviewer and publish or perish software to conduct a computerized bibliometric analysis of the phrase "esports management." the method used descriptivequantitative approach in conjunction with bibliome tric analysis. the data was obtained from the google scholar search results for "esports management." there were 999 articles published between 2017 and 2021, with an increase each year except for 2021 to 2022. this may be demonstrated in 2017 with 58 articles, in 2018 with 92 pieces, in 2019 with 160 articles, in 2020 with 242 articles, and in 2021 with a huge rise to 335 articles. in 2022, however, the number of articles had significantly decreased to 64. based on further findings of this research, it can be concluded that there are several understudied sectors in esports management that may be examined further to increase the efficacy of management in esports. it is anticipated that this research will also serve as an example for further studies in defining and evaluating the research subject, as well as for the esports participants' management team. article history: submitted/received 25 dec 2022 first revised 27 jan 2023 accepted 01 apr 2023 first available online 12 apr 2023 publication date 01 jun 2023 __________________ keywords: bibliometrics, esports management, data analysis, vosviewer. 1. introduction esports is one of the first industries to confront the challenge of transitioning from global to local and online to offline as a whole ecosystem (scholz, 2019; taylor, 2012; scholz, 2020). even though there are still disputes about the nature of esport as a “sport” (franke, 2015; hutchins, 2008; jenny et al., 2017; jonasson & thiborg, 2010; witkowski, international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(1) (2023) 31-48 https://doi.org/10.34010/injiiscom.v4i1.9570 mailto:selvialorena@yahoo.com selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 32 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 2010), the stronger opinion stated that it does can be considered as a sport based on factors such as the presence of enemy, rules and ethics, strategy, as well as winning and defeat (kenzhekanova, 2015). due to the fact that esports is a relatively new industry, it is still evolving as there has been several groundbreaking changes in the recent decades on this industry. since the financial model of the esports industry is unstable, esports organizations prioritize risk management related to future developments such as new markets, franchising, new titles, and the extant fragmentation of the esports industry (kenzhekanova, 2015). this makes it important to properly govern esports as a growing industry (scholz, 2019). however, the research concerning esports is in a fragmented state as there are multiple understudied fields, making it quite hard to research the topic in a thorough manner. conducting a bibliometric study of esports, particularly on the subject of esports management, is one technique to discover such understudied subjects. numerous studies on bibliometric analysis in different domains have been conducted. for instance, digital learning (husaeni & nandiyanto, 2022), computer science (husaeni & nandiyanto, 2023), vocational school (husaeni & nandiyanto, 2023), high school (husaeni & nandiyanto, 2023), covid-19 (hamidah et al., 2020), scientific publications (husaeni et al., 2022; soegoto et al., 2022). additionally, there have been studies on esports, such as those by sousa et al. that addressed the physiological and cognitive functions in competitive esports matches (sousa et al., 2020). chiu et al. undertook a bibliometric study of esports generally as part of their research on bibliometric analysis of esports (chiu et al., 2021). yamanaka et al., büyükbaykal and burak, arwendria, and kurnia all did separate studies on the same subject as chiu et al. (yamanaka et al., 2021; büyükbaykal & burak, 2020; arwendria, 2021; kurnia, 2021). however, no bibliometric examination of esports management has been conducted. based on the issue, this research aims to conduct a bibliometric analysis research regarding esports management. qualitative-descriptive approach was used with literature review as the data collection method. to analyse the data, vosviewer was used to illustrate the connection between the terms as well as to discover the term and publication trend between the year 2017 to 2022. 2. method this study employed quantitative, descriptive, and bibliometric techniques. for this study, we compiled information from a number of google scholar-listed, previously published journals. this is because google scholar is one of the sources of easily accessible journals. using the program publish or perish, we also conducted a literature review on "esports management". it was decided to use publish or perish to extract bibliometric data from study subjects (jenny et al., 2017). moreover, once the data has been saved from the publish or perish utility as a *ris. file, it can be viewed with the vosviewer application. this research uses vosviewer 1.6.17 and publish or perish 8 as its data collection program. in this research, we combed through data and utilized pertinent data to support our claims on esports management. in https://doi.org/10.34010/injiiscom.v4i1.9570 33 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 alignment with the title, keyword, and abstract requirements of the publish or perish program, we retrieve data from google scholar using the term "esports management" 999 data on the study of esports management were acquired. the research papers that were considered were published between 2017 and 2022. the compiled articles are then saved in *.ris format. after that, we create visualizations with the help of the vosviewer program and utilize bibliometric maps to assess trends. from the prepared database source, we map the article data. three categories of mapping information are used by the vosviewer software: network visualization, visualization overlay, and visualization density. additionally, we apply filters to the phrases shown in the mapping representation of vosviewer. (jonasson & thiborg, 2010). 3. results and discussion 3.1. research developments in the field of esports management esports management research into the changing climate, explains how research on the topic of managing esports has evolved between 2017 and 2022 in fig. 1. research on esports management increases every year, starting from 2017 to 2022 except in 2022. this can be proven in 2017 with the number of articles 58, in 2018 increased to 92 articles, in 2019 increased back to 160 articles, in 2020 increased to 242 articles, pada in 2021 experienced a significant increase to 335 articles, and by 2022 research on esports management decreased drastically, the number of publications to 64. we find 999 publications that match the study subject in the publish or perish software's search results. we selected 20 papers from 20 different journals and books with the most citations from this data (table 1). fig. 1. level of research development on esports management https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 34 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 table 1. article data in the field of esports management no authors title year cites refs 1. hamari and sjöblom. what is esports and why do people watch it? 2017 900 (hamari, 2017) 2. jenny et al. virtual (ly) athletes: where esports fit within the definition of “sport” 2017 455 (jenny, 2017) 3. hallmann & giel. esports–competitive sports or recreational activity? 2018 332 (hallmann, 2018) 4. bányai et al. the psychology of esports: a systematic literature review 2019 211 (banyai et al., 2019) 5. reitman et al. esports research: a literature review 2020 193 (reitman et al., 2020) 6. pizzo et al. esport vs sport: a comparison of spectator motives 2018 175 (pizzo, 2018) 7. difrancisco et al. managing the health of the esport athlete: an integrated health management model 2019 126 (difrancisc o et al., 2019) 8. d himmelstein et al. an exploration of mental skills among competitive league of legend players 2021 119 (himmelste in et al., 2021) 9. r li. good luck has fun: the rise of esports 2017 117 (li, 2017) 10. holden et al. the future is now: esports policy considerations and potential litigation 2017 108 (holden et al., 2017) 11. kane & spradley. recognizing esports as a sport 2017 97 (kane & spradley, 2017) 12. ye et al. mastering complex control in moba games with deep reinforcement learning 2020 94 (ye et al., 2020) 13. jenny et al. esports venues: a new sport business opportunity 2018 91 (jenny, 2018) 14. griffiths. the psychosocial impact of professional gambling, professional video gaming & esports 2017 71 (griffiths, 2017) https://doi.org/10.34010/injiiscom.v4i1.9570 35 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 table 1 (continue). article data in the field of esports management no authors title year cites refs 15. freeman & wohn. esports as an emerging research context at chi: diverse perspectives on definitions 2017 70 (freeman & wohn, 2017) 16. ströh. the esports market and esports sponsoring 2017 60 (stroh, 2017) 17. schaeperko et al. the “new” student-athlete: an exploratory examination of scholarship esports players 2017 53 (schaeperko etter, 2017) 18. chung et al. will esports result in a higher prevalence of problematic gaming? a review of the global situation 2019 49 (chung, 2019) 19. ye et al. towards playing full moba games with deep reinforcement learning 2020 44 (ye, 2020) 20. nagorsky & wiemeyer. the structure of performance and training in esports 2020 43 (nagorsky, 2020) twenty papers that meet the requirements for research are listed in table 1. out of the 20 publications that were chosen, the study on esports management has a greatest citation of 900 and a lowest citation of 43. according to table 1, the articles with the most quotations will be published in 2017 and 2022, respectively. the most papers cited between 2017 and 2022 total 900 articles. that there would be 193 articles mentioned in total by 2020. the year with the most quotes includes up to 900 articles. 3.2. visualization esports management topic area using vosviewer visualization al husaeni and nandiyanto claim that the vosviewer software is used in the esports management field since it has a limited amount of relationships (büyükbaykal & burak, 2020). however, in this investigation, vosviewer requires a minimum of three connections. thus, 26 elements in a total of 10 clusters constitute the end result. using analytical mapping and visualization, a study of the atmosphere of the esports industry was conducted: (i) cluster 1 (5 items) esport, esports consumer, esports game, esports research, gaming (see fig. 2). (ii) cluster 2 (4 items) esports competition, esports event, esports tournament, sport management (see fig. 3). https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 36 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 (iii) cluster 3 (4 items) audience, esports community, esports fan, esports sponsorship (see fig. 4). (iv) cluster 4 (3 items) digitalization, sport, strategic management (see fig. 5). (v) cluster 5 (3 items) participant, professional esports player, video game (see fig. 6). (vi) cluster 6 (3 items) athlete, esports athlete, esports organization (see fig. 7). (vii) cluster 7 (1 item) competitive gaming (see fig. 8). (viii) cluster 8 (1 item) development (see fig. 9). (ix) cluster 9 (1 item) gambling (see fig. 10). (x) cluster 10 (1 item) los esport (see fig. 11). cluster 1 is represented with the color red, cluster 2 with the color green, cluster 3 with the color blue-old, cluster 4 with the color yellow. cluster 5 with the color purple, cluster 6 with the color cyan, cluster 7 with the color orange, cluster 8 with the color brown, cluster 9 with the color pink, and cluster 10 with the color draco turquoise. fig. 2. cluster 1 visualization esports management network https://doi.org/10.34010/injiiscom.v4i1.9570 37 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 fig. 3. cluster 2 visualization esports management network fig. 4. cluster 3 visualization esports management network https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 38 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 fig. 5. cluster 4 visualization esports management network fig. 6. cluster 5 visualization esports management network https://doi.org/10.34010/injiiscom.v4i1.9570 39 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 fig. 7. cluster 6 visualization esports management network fig. 8. cluster 7 visualization esports management network https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 40 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 fig. 9. cluster 8 visualization esports management network fig. 10. cluster 9 visualization esports management network https://doi.org/10.34010/injiiscom.v4i1.9570 41 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 fig. 11. cluster 10 visualization esports management network 3.3. network visualization esports management topic area using vosviewer each term's mapping in the vosviewer program is segregated into the first of three categories, visualization network. the representation of a connected network one of the map's features. existing relationship as represented in a network representation, or the line connecting two objects (see fig. 12). visualization network from the item "esports game" obtained using the vosviewer program is shown in fig. 12. each cluster where in each individual area or investigated issue is depicted in fig. 12. the esports management climate, which includes cluster 10 and has a total strength of 40 and occurrence of 43, is seen in fig. 12 above. esports climate connected to cluster 1 (5 items) esport, esports consumer, esports game, esports research, gaming, cluster 2 (4 items) esports competition, esports event, esports tournament, sport management, cluster 3 (4 items) audience, esports community, esports fan, esports sponsorship. cluster 4 (3 items) digitalization, sport, strategic management, cluster 5 (3 items) participant professional, esports play, video game. cluster 6 (3 items) athlete, esports athlete, esports organization cluster 7 (1 item) competitive gaming, cluster 8 (1 item) development, cluster 9 (1 item) gambling, cluster 10 (1 item) los esport. 3.4. overlay visualization of esports management topic area using vosviewer the vosviewer software's second visualization network offers overlaystyle visualization mapping. mapping properly overlay visualization focuses on a fresh research phrase. novelty term or thing in research related to the climate of esports managements shown in fig. 13 https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 42 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 fig. 12. visualization esports management network fig. 13. overlay esports management visualization https://doi.org/10.34010/injiiscom.v4i1.9570 43 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 in the depiction of thing or term type overlay, the popularity of each year can be seen. on visualization overlay, various hues indicate the duration of an extension in a specific period. in this research, the years 2017 to 2022 are considered. more dark colors approach purple, indicating that research on a particular thing or term will be concluded by 2017. in the meantime, the hue is approaching yellow in a lighter fashion. 3.5. density visualization of esports management density visualization is the third and final mapping depiction in the vosviewer software. fig. 14 depicts a visualization of density esports management. the colors that appear in a term can be mapped using this method. if the color that appears becomes paler, then interest in the term is increasing. conversely, if the color is becoming darker or more diminished, the frequency of research on the term is decreasing. yellow color terms are depicted in fig. 14 as having a diameter that is relatively large. these concepts are referred to as emission, esports, esports management, gaming, and development. visualization density about climate esports management research is in the picture above, which means that on the map density showing results analysis use all article regarding esports management in 2017-2022. in fig. 14 is depicted a yellow pattern whose keyword density increases as the circle's diameter increases, indicating that they are more prevalent. if the color on the map fades or blends with the background color green, it indicates that the keyword appears less frequently. fig. 14. visualization density esports management network https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 44 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 4. conclusion the purpose of this study is to evaluate and assess the bibliometric literature on esports management. the keyword "esports management" is used to retrieve data based on a subject area containing keywords, abstracts, and titles. after data processing and filtration, 999 relevant articles were obtained. to generate mapping data, a device soft vos viewer is used. using visualization grid, overlay, and density to map data. based on results in mapping and analysis use vos viewer, obtained that study regarding financial management with the term esports management in 2017-2022 decreased from every year to year. in research this, using method bibliometrics to identify theme main in every field studies before, because important to assess novelty in future research. references al husaeni, d. f., & nandiyanto, a. b. d. (2022). bibliometric using vosviewer with publish or perish (using google scholar data): from step -by-step processing for users to the practical examples in the analysis of digital learni ng articles in pre and post covid-19 pandemic. asean journal of science and engineering, 2(1), 19-46. al husaeni, d. f., & nandiyanto, a. b. d. (2022). mapping visualization analysis of computer science research data in 2017-2021 on the google scholar database with vosviewer. international journal of informatics, information system and computer engineering (injiiscom), 3(1), 1-18. al husaeni, d. f., nandiyanto, a. b. d., & maryanti, r. (2023). bibliometric analysis of educational research in 2017 to 2021 using vosviewer: google scholar indexed research. indonesian journal of teaching in science, 3(1), 1-8. al husaeni, d. f., nandiyanto, a. b. d., & maryanti, r. (2023). bibliometric analysis of educational research in 2017 to 2021 using vosviewer: google scho lar indexed research. indonesian journal of teaching in science, 3(1), 1-8. al husaeni, d. n., & nandiyanto, a. b. d. (2023). bibliometric analysis of high school keyword using vosviewer indexed by google scholar. indonesian journal of educational research and technology, 3(1), 1-12. al husaeni, d. n., & nandiyanto, a. b. d. (2023). bibliometric analysis of high school keyword using vosviewer indexed by google scholar. indonesian journal of educational research and technology, 3(1), 1-12. bányai, f., griffiths, m. d., király, o., & demetrovics, z. (2019). the psychology of esports: a systematic literature review. journal of gambling studies, 35, 351365. https://doi.org/10.34010/injiiscom.v4i1.9570 45 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 büyükbaykal, n. g., & burak, i̇. l. i̇. (2020). e-spor kavramına yönelik araştırmaların bibliyometrik analizi bibliometric analysis of researches for e-sport. uluslararası kültürel ve sosyal araştırmalar dergisi, 6(2), 572-583. chiu, w., fan, t. c. m., nam, s. b., & sun, p. h. (2021). knowledge mapping and sustainable development of esports research: a bibliometric and visualized analysis. sustainability, 13(18), 10354. chung, t., sum, s., chan, m., lai, e., & cheng, n. (2019). will esports result in a higher prevalence of problematic gaming? a review of the global situation. journal of behavioral addictions, 8(3), 384-394. difrancisco-donoghue, j., balentine, j., schmidt, g., & zwibel, h. (2019). managing the health of the esport athlete: an integrated health management model. bmj open sport & exercise medicine, 5(1), e000467. franke, t. (2013). the perception of esports-mainstream culture, real sport and marketisation. esports yearbook, 14, 111-144. freeman, g., & wohn, d. y. (2017, may). esports as an emerging research context at chi: diverse perspectives on definitions. in proceedings of the 2017 chi conference extended abstracts on human factors in computing systems (pp. 16011608). griffiths, m. d. (2017). the psychosocial impact of professional gambling, professional video gaming & esports. casino & gaming international, 28, 59-63. hallmann, k., & giel, t. (2018). esports–competitive sports or recreational activity?. sport management review, 21(1), 14-20. hamari, j., & sjöblom, m. (2017). what is esports and why do people watch it?. internet research, 27(2), 211-232. hamidah, i., sriyono, s., & hudha, m. n. (2020). a bibliometric analysis of covid -19 research using vosviewer. indonesian journal of science and technology, 34-41. himmelstein, d., liu, y., & shapiro, j. l. (2021). an exploration of mental skills among competitive league of legend players. in research anthology on rehabilitation practices and therapy (1607-1629). igi global. holden, j. t., kaburakis, a., & rodenberg, r. (2017). the future is now: esports policy considerations and potential litigation. j. legal aspects sport, 27, 46. hutchins, b. (2008). signs of meta-change in second modernity: the growth of e-sport and the world cyber games. new media & society, 10(6), 851-869. jenny, s. e., keiper, m. c., taylor, b. j., williams, d. p., gawrysiak, j., manning, r. d., & tutka, p. m. (2018). esports venues: a new sport business opportunity. journal of applied sport management, 10(1), 8. https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 46 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 jenny, s. e., manning, r. d., keiper, m. c., & olrich, t. w. (2017). virtual (ly) athletes: where esports fit within the definition of “sport”. quest, 69(1), 1-18. jenny, s. e., manning, r. d., keiper, m. c., & olrich, t. w. (2017). virtual (ly) athletes: where esports fit within the definition of “sport”. quest, 69(1), 1-18. jonasson, k., & thiborg, j. (2010). electronic sport and its impact on future sport. sport in society, 13(2), 287-299. kane, d., & spradley, b. d. (2017). recognizing esports as a sport. the sport journal, 19(5). kenzhekanova, k. k. (2015). linguistic features of political discourse. mediterranean journal of social sciences, 6(6 s2), 192. kurnia, s. (2021). science, technology, engineering, art and mathematics (steam) di pendidikan sains: analisis bibliometrik dan pemetaan literatur penelitia n menggunakan perangkat lunak vosviewer (doctoral dissertation, uin raden intan lampung). li, r. (2017). good luck have fun: the rise of esports. simon and schuster. masruroh, b., laksana, e. p., rosyida, f., harianti, l. r., & maysa, f. (2022). analisis sitasi jurnal pendidikan geografi: kajian, teori, dan praktik dalam bidang pendidikan dan ilmu geografi periode 2019-2021. jurnal integrasi dan harmoni inovatif ilmu-ilmu sosial (jihi3s), 2(3), 204-209. mulyawati, i. b., & ramadhan, d. f. (2021). bibliometric and visualized analysis of scientific publications on geotechnics fields. asean journal of science and engineering education, 1(1), 37-46. nagorsky, e., & wiemeyer, j. (2020). the structure of performance and training in esports. plos one, 15(8), e0237584. nandiyanto, a. b. d., al husaeni, d. n., & al husaeni, d. f. (2021). a bibliometric analysis of chemical engineering research using vosviewer and its correlation with covid-19 pandemic condition. journal of engineering science and technology, 16(6), 4414-4422. nandiyanto, a. b. d., girsang, g. c. s., maryanti, r., ragadhita, r., anggraeni, s., fauzi, f. m., ... & al-obaidi, a. s. m. (2020). isotherm adsorption characteristics of carbon microparticles prepared from pineapple peel waste. communications in science and technology, 5(1), 31-39. pizzo, a., baker, b., na, s., lee, m., kim, d., & funk, d. (2018). esport vs sport: a comparison of spectator motives. faculty/researcher works. https://doi.org/10.34010/injiiscom.v4i1.9570 47 | international journal of informatics information system and computer engineering 4(1) (2023) 31-48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 ragadhita, r., & nandiyanto, a. b. d. (2022). computational bibliometric analysis on publication of techno-economic education. indonesian journal of multidiciplinary research, 2(1), 213-220. reitman, j. g., anderson-coto, m. j., wu, m., lee, j. s., & steinkuehler, c. (2020). esports research: a literature review. games and culture, 15(1), 32-50. schaeperkoetter, c. c., mays, j., hyland, s. t., wilkerson, z., oja, b., krueger, k., ... & bass, j. r. (2017). the “new” student-athlete: an exploratory examination of scholarship esports players. journal of intercollegiate sport, 10(1), 1-21. scholz, t. m. (2020). deciphering the world of esports. international journal on media management, 22(1), 1-12. scholz, t. m., & scholz, t. m. (2019). a short history of esports and management. esports is business: management in the world of competitive gaming, 17-41. scholz, t. m., & scholz, t. m. (2019). introduction: the emergence of esports. esports is business: management in the world of competitive gaming, 1-16. soegoto, h., soegoto, e. s., luckyardi, s., & rafdhi , a. a. (2022). a bibliometric analysis of management bioenergy research using vosviewer application. indonesian journal of science and technology, 7(1), 89-104. sousa, a., ahmad, s. l., hassan, t., yuen, k., douris, p., zwibel, h., & difrancisco donoghue, j. (2020). physiological and cognitive functions following a discrete session of competitive esports gaming. frontiers in psychology, 11, 1030. ströh, j. h. a. (2017). the esports market and esports sponsoring. tectum wissenschaftsverlag. taylor, t. l. (2012). raising the stakes: e-sports and the professionalization of computer gaming. mit press. witkowski, e. (2009). probing the sportiness of esports. esports yearbook. norderstedt: books on demand gmbh, 53-56. yamanaka, g. k., campos, m. v., roble, o. j., & mazzei, l. c. (2021). esport: a stateof-the-art review based on bibliometric analysis. journal of physical education and sport, 21(6), 3547-3555. ye, d., chen, g., zhang, w., chen, s., yuan, b., liu, b., ... & liu, w. (2020). towards playing full moba games with deep reinforcement learning. advances in neural information processing systems, 33, 621-632. https://doi.org/10.34010/injiiscom.v4i1.9570 selvia lorena br ginting. a computational bibliometric analysis of esport management using… | 48 doi: https://doi.org/10.34010/injiiscom.v4i1.9570 p-issn 2810-0670 e-issn 2775-5584 ye, d., liu, z., sun, m., shi, b., zhao, p., wu, h., ... & huang, l. (2020, april). mastering complex control in moba games with deep reinforcement learning. in proceedings of the aaai conference on artificial intelligence, 34(04), 6672-6679. https://doi.org/10.34010/injiiscom.v4i1.9570 1 | international journal of informatics information system and computer engineering 4(1) (2023) 1-10 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 ict-based literacy evaluation in nigeria educational sector: case study in kwara state hammed olalekan bolaji*, tinuke bilikis ibrahim-raji 1department of science education, faculty of education, al -hikmah university, ilorin-nigeria 2department of educational management and counseling, faculty of education, al-hikmah university, ilorin-nigeria *corresponding email: atinukebilikis259@gmail.com a b s t r a c t s a r t i c l e i n f o ict is now a necessity for both professionals and organizations due to its pervasiveness across all fields of human endeavour. the literacy skills level plays a major role in its application for routine responsibilities and the pace at which task is complemented. for efficient service delivery in the public service, this study evaluated the ict literacy abilities and their application among the staff of education agencies. it used a descriptive cross-sectional survey design. structured items on the ict skills assessment and utilization questionnaire (ictsauq) were administered to fifty staff using convenient sampling techniques. to respond to the research questions posed by this study, descriptive and inferential statistics were used. the question was addressed using percentage means and standard deviations, and the questions were analysed using a t-test. the results showed that having a basic understanding of ict helps do administrative tasks daily. however, the staff of the education agencies lacked the necessities for their daily routine of managerial responsibilities and operations. hence, it was suggested that staff of the education agencies in kwara state must be exposed to the required ict skills to perform the routine functions at the optimal level. additionally, it was suggested that agency staff members be encouraged to consistently improve their ict literacy abilities through self-training and group work to improve the competence of service delivery in the educational sector. article history: submitted/received 18 jan 2023 first revised 20 feb 2023 accepted 25 march 2023 first available online 1 april 2023 publication date 01 june 2023 aug 2018 __________________ keywords: ict, utilization, literacy skills. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(1) (2023) 1-10 https://doi.org/10.34010/injiiscom.v4i1.9274 mailto:atinukebilikis259@gmail.com bolaji et al., ictbased literacy evaluation in nigeria…| 2 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 1. introduction the level of literacy needed to operate computers and other related technologies effectively is known as information and communication technology (ict) ability (ukachi, 2015). a grasp of how to use common software systems and platforms and a practical knowledge of computer programming and its applications could be on the skill spectrum (alemu, 2015). the use of ict in educational practice is widely regarded as an empowering tool that encourages change and stimulates the development of 21st-century skills (bolaji & adeoye, 2022). currently, the pervasiveness of ict in every sphere of human endeavour cannot be overemphasized, and education is also experiencing the feeling of it in its routine activities and practices. this requires some level of literacy skill in the use of ict to perform one function or the other among educational administrators. the most crucial worldwide resource for selfactualization is ict with its corresponding literacy skills (olatoye, 2019). ict as a form of technological development can improve economic prospects, and enhance governance and service delivery for the socio-economic development of a society (ogwu & ogwu, 2015). it is on this premise that emwanta (2013) advised that ict literacy skills and abilities should be acquired to maximize its potential for educational practices (emwanta, 2013). the advent of numerous ict tools has resulted in significant changes in the educational system around the world. the provision of these amenities by government agencies in the workplace has been shown to improve service delivery efficiency (adeleke, 2016) and should be sustained with friendly ict policies. the federal government of nigeria created an ict policy in 2001 after considering its advantages and the national information technology development agency (nitda) was founded as a result of the policy in which its goal includes ensuring that ict resources are easily accessible to support effective national development and integrate it into the civil service, notably the education sector (nitda, 2017). furthermore, emwanta (2013) posited that policy is a vital instrument for the promotion and sustenance of national development (emwanta, 2013). education policy is well recognized for its ability to transmit desirable values like work ethics, loyalty, integrity and justice; all of which are necessary for individual survival and societal growth. education could be the most crucial tool for change, and staffs, irrespective of gender are the driving force behind it. the gender of an individual always plays a moderating influence on routine activities, and the use of ict is not an exception. gender influences the use of ict for teaching as revealed in the study conducted by irfan et al (2014) where it was found that male teachers frequently use ict in comparison to female colleagues, and it as well extends to creating presentation materials for instructional delivery (guillen et al., 2019). important also, it was further revealed that gender didn’t have a major influence on the use of ict for information seeking which is at variance with the initial finding. therefore, it can be concluded that gender swings its influence on individuality and the responsibilities routinely performed. gender might not necessarily influence every variable of an individual, and as reported by egunjobi and fabunmi https://doi.org/10.34010/injiiscom.v4i1.9274 3 | international journal of informatics information system and computer engineering 4(1) (2023) 1-10 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 (2017), a relationship did not exist between competence and gender (olusegun & adesoji, 2017). in addition, gender didn’t have any moderating influence on the use of ict for information retrieval and its actual use (durante, 2013). the attitude of an individual while using ict might also not influence routine activities (adenuga et al., 2011). whereas it was discovered that gender influences the use of ict for accessing social networks in females was found to be more prevalent in the use of it for communications (hilbert, 2011). however, males are found to possess more skills in the use of ict (van dijk, 2015). a further advancement is the ict literacy level perceived to be higher in female students but does not influence their operational skills over their male counterparts (zhou et al., 2014). the primary goal of this research was to examine the degree and scope of ict literacy use among education agency workers. in particular, the study: evaluated the degree of ict literacy among kwara state's educational agency staff. examined how the staff of education agencies in kwara state used their knowledge of ict. figured out how to get education agencies' workers in kwara state to become ict literate. examined the difficulties in teaching ict literacy to the workers of kwara state's educational agencies. the purpose of the study is what is the degree of ict literacy among kwara state's educational agency staff? do the kwara state employees of education agencies possess ict literacy skills? how are ict literacy skills acquired by the kwara state education agency staff? what are the difficulties that the kwara state workers of education agencies face while using their ict literacy skills? 2. method this study is descriptive of the crosssectional survey type. the instrument for this study was a researcher-designed questionnaire titled ict skills assessment and utilization questionnaire (ictsauq) comprising a close-ended questionnaire for the collection of data on the ict literacy skill and its utilization among the staff of education agencies in kwara state. the sampling technique employed is a non-probability convenient sample and the technique was adopted because the researcher is specific about the participants who are staff working in the agencies relating to education within the civil service. the agencies randomly selected for this study are the teaching service commission, state universal basic education board, scholarship board and mass literacy agency. hence, the sample size for this was 50 and 10 of each of the staff were conveniently selected across the education agencies in kwara state. the questionnaire has three sections of six items each to measure the variables under study. the five–point likert is considered as an internal scale for all the questions statements. if the mean is from 1 to 1.8 it signifies strongly disagree and for the mean from 1.81 to 2.60, it signifies disagree. likewise, if the mean is from 2.61 to 3.40 indicate undecided and the mean from 3.41 to 4.20 signifies agree. also, from 4.21 to 5 the mean is strongly agreed. the questionnaire was validated by three educators comprising an educational technologist, educational manager and computer educator and thereafter subjected to a reliability test using cronbach alpha which yielded https://doi.org/10.34010/injiiscom.v4i1.9274 bolaji et al., ictbased literacy evaluation in nigeria…| 4 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 0.87. the responses were collected manually and subjected to both descriptive and inferential statistical analysis. the research questions were answered using mean and standard deviation while the hypothesis was analyzed using a t-test. 3. results and discussion across all the education agencies, it is evident that there are more females than males. for instance, the total average of females overall is 60% which translated to 30 out of 50 respondents examined for this study. this implies that females dominated the staff strength of education agencies in kwara state and it laid the foundation for the need to give gender attention regarding the variables under study (see tables 1 and 2). question 1: what is the level of ict literacy skills among the staff of education agencies in kwara state? table 1. demographics characteristic of the respondents mdas (ministries, departments and agencies) male female total teaching service commission 6(40%) 9 (60%) 15 state universal basic education board 5 (33.33%) 10 (66.6%) 15 scholarship board 5 (50%) 5 (50%) 10 mass literacy agency 4 (40%) 6 (60%) 10 total 20 (40%) 30 (60%) 50 table 2. mean and standard deviation on ict literacy skills among staff education agencies statements n mean std. deviation i can create a multimedia presentation (with sound, pictures and video). 50 3.26 1.192 i can copy files from one location into another location. 50 3.42 1.052 i can send and open an attachment from an email, using a computer email program. 50 3.08 1.482 i can use the world wide web address to find useful information 50 3.84 0.650 i can use search engines to search for information e.g, yahoo, google, and youtube. i can use the internet and its various features 50 50 3.00 3.86 1.262 0.808 https://doi.org/10.34010/injiiscom.v4i1.9274 5 | international journal of informatics information system and computer engineering 4(1) (2023) 1-10 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 from table 2, in the first statement, the mean is 3.26. hence, this shows that the majority of participants cannot decide whether they can create a multimedia presentation (with sound, pictures and video) or not. the means of the second statement is, 3.42; it means that the majority of participants agreed to have the ability to copy files from one location into another location. the means of the third statement is, 3.08; it means that the majority of participants cannot decide whether they can send and open an attachment from an email, using a computer email program or not. the means of the fourth statement is, 3.84; it means that the majority of participants agreed that they can use www address to find useful information. the means of the fifth statement is 3.00, which means that the majority of participants cannot decide whether they can use search engines to search for information e.g, yahoo, google, and youtube or not. the means of the sixth statement is, 3.86; it means that the majority of participants strongly agreed that they can use the internet and its various features (see table 3). question 2: do the staff of education agencies in kwara state utilize ict literacy skills? table 3 shows that the mean of the first statement is 3.50, this shows that the majority of participants agreed that they can start up and shut down the computer system. the means of the second statement is, 3.66; it means that the majority of participants agreed to have the ability to open, create, edit, backup, save and delete documents or files on the computer. the means of the third statement is, 3.20; it means that the majority of participants cannot decide whether they can copy a file from a floppy disk or flash drive (usb). the means of the fourth statement is, 3.54; it means that the majority of participants agreed that they can use microsoft word for typing. the means of the fifth statement is, 3.24; it means that the majority of participants cannot decide whether they can use microsoft excel for analysis or not. the means of the sixth statement is, 3.40; it means that the majority of participants cannot decide whether they can use microsoft powerpoint for presentation (table 4). question 3: how does the staff of education agencies in kwara state acquire ict literacy skills table 4 indicates that the mean of the first statement is, 2.28; this showing that the majority of participants disagreed with the statement that they went to seminars and conferences. the means of the second statement is, 3.72; it means that the majority of participants agreed to do personal and self-training. the means of the third statement is, 3.24; it means that the majority of participants cannot decide whether they got ict literacy from colleagues in the office. the means of the fourth statement is 2.30, which means that the majority of participants disagreed that they got ict literacy skills through government training. the means of the fifth statement is, 3.32; it means that the majority of participants cannot decide whether they acquire ict literacy skills from in-house training (see table 5). question 4: what are the challenges of literacy skill use of ict among the staff of education agencies in kwara state? https://doi.org/10.34010/injiiscom.v4i1.9274 bolaji et al., ictbased literacy evaluation in nigeria…| 6 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 table 3. mean and standard deviation on the utilization of ict literacy skills among the staff of the education agencies table 4. mean and standard deviation on the acquisition of ict literacy skills statements n mean std. deviation i went to seminars and conferences 50 2.28 1.070 i did personal and self-training 50 3.72 0.834 through colleagues in the office 50 3.24 0.938 through government training 50 2.30 1.233 in-house training 50 3.32 1.377 table 5. mean and standard deviation on the challenges of ict literacy skills using statements n mean std. deviation the use of productive e-resources may be constrained by inadequate ict skills. 50 3.18 1.366 the usage of electronic information resources might be hampered by inadequate ict literacy abilities. 50 3.32 1.168 electronic information resources are ineffectively used when ict skills are lacking. 50 3.76 1.393 the use of e-information resources is hindered by the inability to operate a computer. 50 3.36 1.411 access to electronic resources might be badly impacted by a lack of computer skills. 50 3.36 1.453 statements n mean std. deviation i can start up and shut down a computer system. 50 3.50 1.568 i can open, create, edit, backup, save and delete documents or files on the computer 50 3.66 1.136 i can copy a file from a floppy disk or flash drive (usb) 50 3.20 1.278 i can use microsoft word for typing 50 3.54 0.862 i can use microsoft excel for analysis 50 3.24 1.519 i can use microsoft powerpoint for presentation 50 3.40 0.928 https://doi.org/10.34010/injiiscom.v4i1.9274 7 | international journal of informatics information system and computer engineering 4(1) (2023) 1-10 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 table 5 the mean of the first statement is, 3.18; this shows that the majority of participants agreed on the use of productive e-resources may be constrained by inadequate ict skills. the means of the second statement is, 3.32; it means that the majority of participants agreed on the usage of electronic information resources might be hampered by inadequate ict literacy abilities. the means of the third statement is, 3.76; it means that the majority of participants agreed that electronic information resources are ineffectively used when ict skills are lacking. the means of the fourth statement is, 3.36; it means that the majority of participants cannot decide whether the use of einformation resources is hindered by the inability to operate a computer. the means of the fifth statement is, 3.36; it means that the majority of participants cannot decide whether access to electronic resources might be badly impacted by a lack of computer skills. 4. discussion the results of table two showed that the analysis of multimedia presentations, search engine usage, and internet usage showed that skills are essential for the efficient use of ict. this corroborates the finding from balarabe's (2020) study that pupils have a basic comprehension of ict capabilities, including competency with ms word, ms powerpoint, internet searching, and other related abilities (yushau, 2020). because it is the driving force behind ict abilities. the study of table three's data revealed that certain employees have trouble starting up and shutting down computers and are less conversant with presentation-related microsoft programs like word, excel, and powerpoint. when compared to all respondents, the percentage is minuscule. this could signal that the staff is making progress in their pursuit of ict and that they recognize the need to continually improve their ict skills to be able to meet the demands of each cadre for ict usage. table 4 results showed how education agencies' staff members develop their ict skills. they did, however, gain it mainly through personal and self-training. the results of oluwayemi et al. (2021), who testified that ict literacy abilities are more often used in training, concurred with this (olatoye, 2021). the outcome from table five demonstrates that staff members do have difficulties due to a lack of ict skills and expertise. however, staff members claimed to have trouble using electronic information resources, which may be caused by a lack of ict literacy. this is to the results of makori (2016), who asserted that it is dificult to give pupils the ict skills and resources they need (makori, 2016). according to makhmudov, k., shorakhmetov, s., & murodkosimov, a. (2020), not all subject teachers need to be experts at using computers, even if computer literacy is a must. for their teachings to be more effective and to better serve their students, they should possess a certain set of skills. these skills include the following (makhmudov et al., 2020). a) the ability to read and write simple computer programs; b) the ability to use computer programs and documentation that are educational in nature; c) the ability to use computer terminology, especially as it relates to hardware; d) the ability to identify educational problems that can and cannot be solved using the computer; e) the https://doi.org/10.34010/injiiscom.v4i1.9274 bolaji et al., ictbased literacy evaluation in nigeria…| 8 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 ability to locate information on computing as it relates to education; and f) the ability to discuss the moral and human-impact issues. koltay (2011) made the case that information literacy is crucial for the growth of democracy, cultural participation, and active civic participation (koltay, 2011). knowledge workers who heavily rely on the internet and computer tools are especially in need of this literacy. information literacy also places a strong emphasis on the necessity of recovering and careful selection of the information available in the workplace, in education, and in other areas of individual decision-making, particularly in the domains of citizenship and health. information literacy training places a strong emphasis on the critical thinking, metacognitive, and procedural skills needed to find information in particular fields, settings, and contexts. the acknowledgement of the message of quality, authenticity, and credibility is prioritized (hobbs, 2006). in a study on how information literacy is seen in educational environments, the workplace, and the community, lloyd and williamson (2008) came to the conclusion that the context is a significant component in shaping the phenomena (lloyd, 2008). information literacy, according to catts and lau (2008), is appropriate in all areas of human development and is defined as the capacity to recognize information needs, evaluate their quality, manage this information, use it effectively, and do so in an ethical manner, in addition to the capacity to produce and share the knowledge attained through the application of information (catts & lau, 2008). there are several common elements among the definitions given, with perhaps the most significant one being the understanding that information skills cannot be seen in isolation since they are interrelated processes that entail how people think about and use information (eisenberg et al., 2004). combination of information and computer literacy (ict literacy) has been explained by oecd (organisation for economic co-operation and development) and by (santos et al., 2019). as the interest, attitude and ability of individuals to properly use digital technology and communication tools to access, manage, integrate and evaluate information, construct new knowledge, and communicate with others in order to effectively participate in society as shown in table 5. 5. conclusion from the findings, ict utilization is evident in all the departments and agencies in the education ministries of kwara state. conclusively, education agency staff possess ict skills that are useful for their profession through selftraining. however, gender has no significant influence on the ict skills of the staff. the need for effective utilization of ict devices within the agencies in the state education sector calls for prompt action by all relevant stakeholders to meet the present global technology challenges. hence the need to tackle ict deficiency among the staff of the agencies require essential attention. government should ensure that staff development and ict utilization should be prioritized and funds should be duly allocated for it in the education sector. https://doi.org/10.34010/injiiscom.v4i1.9274 9 | international journal of informatics information system and computer engineering 4(1) (2023) 1-10 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 references adeleke, d. s., & emeahara, e. n. (2016). relationship between information literacy and use of electronic information resources by postgraduate students of the university of ibadan. library philosophy and practice, 1. adenuga, r. a., owoyele, j. w., & adenuga, f. t. (2011). gender and socio -economic background differentials in students’ attitude to information and communication technology education in nigerian secondary schools: implications for policy, ict education and counselling. international journal of psychology and counseling, 3(9), 162-166. alemu, b. m. (2015). integrating ict into teaching-learning practices: promise, challenges and future directions of higher educational institutes. universal journal of educational research, 3(3), 170-189. bolaji, h. o., & adeoye, m. a. (2022). accessibility, usability, and readiness towards ict tools for monitoring educational practice in secondary schools in ilorin metropolis. indonesian journal of multidiciplinary research, 2(2), 257-264. catts, r. and lau, j. (2008). towards information literacy indicators. technical report, unesco eisenberg, m. b., lowe, c. a., and spitzer, k. l. (2004). information literacy: essential skills for the information age. greenwood publishing group, 88 post road west, westport. durante, f., fiske, s. t., kervyn, n., cuddy, a. j., akande, a., adetoun, b. e., ... & storari, c. c. (2013). nations' income inequality predicts ambivalence in stereotype content: how societies mind the gap. british journal of social psychology, 52(4), 726-746. emwanta, m., & nwalo, k. i. n. (2013). influence of computer literacy and subject background on use of electronic resources by undergraduate students in universities in south-western nigeria. international journal of library and information science, 5(2), 29-42. guillén-gámez, f. d., lugones, a., & mayorga-fernández, m. j. (2019). ict use by pre-service foreign languages teachers according to gender, age and motivation. cogent education, 6(1), 1574693. hilbert, m. (2011, november). digital gender divide or technologically empowered women in developing countries? a typical case of lies, damned lies, and statistics. in women's studies international forum, 34(6), 479-489. pergamon. hobbs, r. (2006). multiple visions of multimedia literacy: emerging areas of synthesis. international handbook of literacy and technology, 2, 15-28. koltay, t. (2011). the media and the literacies: media literacy, information literacy, digital literacy. media, culture & society, 33(2), 211-221. lloyd, a., & williamson, k. (2008). towards an understanding of information literacy in context: implications for research. journal of librarianship and information science, 40(1), 3-12. https://doi.org/10.34010/injiiscom.v4i1.9274 bolaji et al., ictbased literacy evaluation in nigeria…| 10 doi: https://doi.org/10.34010/injiiscom.v4i1.9274 p-issn 2810-0670 e-issn 2775-5584 makhmudov, k., shorakhmetov, s., & murodkosimov, a. (2020). computer literacy is a tool to the system of innovative cluster of pedagogical education. european journal of research and reflection in educational sciences, 8(5). makori, e. o., & mauti, n. o. (2016). digital technology acceptance in transformation of university libraries and higher education institutions in kenya. national information technology development agency (nitda). (2017, february 2). nitda inaugurates cobit 5 national implementation committee. ogwu, e. n., & ogwu, f. c. (2016). comparative analysis of microsoft package (msp) competence among teacher trainee students in botswana and nigeria: implications for curriculum practices. olatoye, o. i. (2019). ict literacy skills and demographic factors as determinants of electronic resources use among the undergraduate students in the selected universities the eastern cape, south africa. olatoye, o. i., nekhwevha, f., & muchaonyerwa, n. (2021). ict literacy skills proficiency and experience on the use of electronic resources amongst undergraduate students in selected eastern cape universities, south africa. library management, 42(6/7), 471-479. olusegun, e. a., & adesoji, f. f. (2017). gender influence of ict competence of undergraduates in state–owned universities in the south–west nigeria. journal of applied information science and technology, 10(1), 140-151. santos, g. m., ramos, e. m., escola, j., & reis, m. j. (2019). ict literacy and school performance. turkish online journal of educational technology-tojet, 18(2), 19-39. ukachi, n. b. (2015). information literacy of students as a correlate of their use of electronic resources in university libraries in nigeria. the electronic library, 33(3), 486-501. van dijk, c. (2015). siektes en sindrome geassosieer met'n hoë reënval -dr chris se notas. veeplaas, 6(1), 93. yushau, b., & nannim, f. a. (2020). investigation into the utilization of ict facilities for teaching purposes among university lecturers: influence of gender, age, qualification and years of teaching experience. pedagogical research, 5(2). zhou, j., chu, h., li, c., wong, b. h. y., cheng, z. s., poon, v. k. m., ... & yuen, k. y. (2014). active replication of middle east respiratory syndrome coronavirus and aberrant induction of inflammatory cytokines and chemokines in human macrophages: implications for pathogenesis. the journal of infectious diseases, 209(9), 1331-1342. https://doi.org/10.34010/injiiscom.v4i1.9274 1 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 mapping visualization analysis of computer science research data in 2017-2021 on the google scholar database with vosviewer dwi fitria al husaeni, asep bayu dani nandiyanto universitas pendidikan indonesia, indonesia e-mail: dwifitriaalhusaeni@upi.edu a b s t r a c t s a r t i c l e i n f o the purpose of this research is to examine the development and interrelationships between terms in computer science research using mapping analysis with vosviewer. the research data was collected from the google scholar database for the period 2017-2021 using the publish or perish 7 application. data collection was based on the keyword "computer science". the data search results found 992 articles that were considered relevant. the results showed that computer science research experienced high popularity in 2018 with a total of 232 articles. computer science research experienced a decline in research in 2019-2021. based on the mapping analysis that has been done using the vosviewer application, computer science terms are connected to 4 main terms in each cluster, namely student, computer science education, education, and skills. computer science research is mostly associated with the term student, namely the strength of link 221. this research can be used as a reference in determining the research theme or research discussion topic in the field of computer science. article history: received 5 may 2022 revised 15 may 2022 accepted 25 may 2022 available online 26 june 2022 ___________________ keywords: bibliometric, computer science, mapping analysis, vosviewer 1. introduction computer science is generally defined as the study of computers, hardware, and software (armoni & gal-ezer, 2014). computer science is rooted in electronics, mathematics, and linguistics (alhazov, 2010). computer science covers a wide range of computer-related topics, from abstract analysis of algorithms to more specific topics such as programming languages, software, and hardware. computer science focuses more on computer programming and software engineering. computer science is a international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 1-18 al husaeni et al. mapping visualization analysis of computer science… | 2 branch of science that deals with computers and computing. computer science covers theory, component testing, and design and includes questions related to the theoretical understanding of computer devices, programs, and systems, experimental development and testing of computational concepts, design methodologies, algorithms, and tools to achieve them, analytical methods to demonstrate that implementation conforms to requirements. the development of the times is getting faster due to drastic technological changes. therefore, computer science personnel are needed in the workplace because all human needs can be facilitated because of technology (tussyadiah, 2020). therefore, research on computer science also needs to know its development so that it can continue to develop as the times. analysis of mapping visualization can be used to determine the development of computer science research. currently, there are many studies on mapping analysis, including new science mapping analysis software tool (cobo et al., 2012), science mapping analysis using r-tool (aria & cuccurullo, 2017), equity mapping analysis (wolch et al., 2005; jurado de los santos et al., 2020; talen, 1998; mennis & jordan, 2005), mapping analysis about magnetic properties and energy (xiang et al., 2013), and mapping analysis of the pipeline for pooled rna-seq (hill et al., 2013). however, there is no research on mapping analysis that discusses research in the field of computer science based on text data and bibliographic data using vosviewer. therefore, this research examines the analysis of mapping data for computer science publications using vosviewer by visualizing the mapping into three types, namely network visualization, overlay visualization, and density visualization. thus, through this research, it can be seen that the terms of computer science research are connected to facilitate the search for other fields of discussion that have high novelty in the field of computer science. 2. methodology this study uses a mapping analysis method on a data set of articles published in journals from 2017 to 2021 indexed by google scholar. data retrieval from the google scholar database is open source. to get the data from the research, we use the reference manager application publish or perish 7. all data were obtained on 12 january 2022. the publish or perish software review the literature on a predefined topic is "computer science". detailed information for installing and using software (google scholar and publish or perish 7) and a step-by-step process for obtaining data were described in our previous study (al husaeni & nandiyanto, 2022). there are several stages carried out in this research: (i) determination of study topics, (ii) collection of publication data is taken from the google scholar database using the publish or perish 7 application. (iii) processing of text data and bibliometric data on articles that have been obtained using microsoft excel application, which is converted into three file formats, namely research information systems (.ris), comma-separated value format (*.csv) and excel workbook (*.xlsx) 3 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 (iv) visualization of publication data mapping using the vosviewer application version 1.6.16, and (v) analysis of mapping analysis results. visualization of mapping text data and article bibliometric data is made in 3 types, namely network visualization, density visualization, and overlay visualization based on the relationship between existing items. data mapping is carried out in 2 steps, namely mapping based on text data and mapping based on bibliographic data. data mapping based on text data found 5972 terms. the terms that have been found are re-sorted with several provisions including the minimum number of occurrences of a term is 10. therefore, the number of terms used in the mapping analysis is 159 terms. data mapping based on text data is used to see the relationship between existing terms and is used in research in the field of computer science. the second data mapping on the same data was carried out based on bibliographic data. this mapping was carried out to find connections and also to see authors who contributed quite high to research in the field of computer science as recorded by google scholar. the rules used in making this data mapping include the maximum number of authors per document is 25 authors, the minimum number of documents of an author is 3 times. thus, it was found that 87 authors from 2073 authors met the criteria and entered the data mapping process. 3. result and discussion 3.1. publication data search results the search results for published data on computer science found 992 articles in the google scholar database for 2017-2021. table 1 presents one of the article data used in the vosviewer mapping analysis. all article data that has been obtained are then sorted based on their citation values so that the 20 best articles with the highest citations are found as presented in table 1. from the data in table 1, it is found that the highest citations were dominated by articles published in 2017, which were 20 articles with an average number of citations of 114.4 times. the average citation in 2017 for 20 articles with the highest citations was 22.88 times per year. table 1. computer science publication data no authors title year cites cites per year cites per author refs 1 weintrop, d., & wilensky, u comparing block-based and textbased programmin g in high school computer science classrooms 2017 204 40.80 102 weintrop & wilensky (2017) al husaeni et al. mapping visualization analysis of computer science… | 4 no authors title year cites cites per year cites per author refs 2 webb, m., davis, n., bell, t., katz, y. j., reynolds, n., chambers, d. p., & sysło, m. m. computer science in k12 school curricula of the 2lst century: why, what, and when? 2017 182 36.40 30 webb et al. (2017) 3 borrego, c., fernández, c., blanes, i., & robles, s. room escapes at class: escape games activities to facilitate the motivation and learning in computer science 2017 182 36.40 46 borrego et al. (2017) 4 sax, l. j., lehman, k. j., jacobs, j. a., kanny, m. a., lim, g., monjepaulson, l., & zimmerman, h. b. anatomy of an enduring gender gap: the evolution of women's participation in computer science 2017 177 35.40 35 sax et al. (2017) 5 wang, d., liang, y., xu, d., feng, x., & guan, r. a contentbased recommender system for computer science publications 2018 173 43.25 35 wang et al. (2018) 6 vakil, s. ethics, identity, and political vision: toward a justicecentered approach to 2018 114 28.50 114 vakil (2018) 5 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 no authors title year cites cites per year cites per author refs equity in computer science education 7 passey, d. computer science (cs) in the compulsory education curriculum: implications for future research 2017 95 19.00 95 passey (2017) 8 giannakos, m. n., pappas, i. o., jaccheri, l., & sampson, d. g. understandin g student retention in computer science education: the role of environment, gains, barriers, and usefulness 2017 87 17.40 22 giannako s et al. (2017) 9 garcia, r., falkner, k., & vivian, r. systematic literature review: selfregulated learning strategies using elearning tools for computer science 2018 80 20.00 27 garcia et al. (2018) 10 leytonbrown, k., milgrom, p., & segal, i. (2017). economics and computer science of a radio spectrum reallocation 2017 66 13.20 22 leytonbrown et al. (2017) al husaeni et al. mapping visualization analysis of computer science… | 6 no authors title year cites cites per year cites per author refs 11 fields, d. a., kafai, y., nakajima, t., goode, j., & margolis, j. (2018). putting making into high school computer science classrooms: promoting equity in teaching and learning with electronic textiles in exploring computer science 2018 65 16.25 13 fields et al. (2018) 12 qian, y., hambrusch, s., yadav, a., & gretter, s. who needs what: recommenda tions for designing effective online professional development for computer science teachers 2018 61 15.25 15 qian et al. (2018) 13 weintrop, d. block-based programmin g in computer science education 2019 56 18.67 56 weintrop (2019) 14 bonham, k. s., & stefan, m. i. women are underreprese nted in computationa l biology: an analysis of the scholarly literature in biology, 2017 55 11.00 28 bonham & stefan (2017) 7 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 no authors title year cites cites per year cites per author refs computer science, and computationa l biology 15 burnette, j. l., hoyt, c. l., russell, v. m., lawson, b., dweck, c. s., & finkel, e. a growth mindset intervention improves interest but not academic performance in the field of computer science 2020 53 26.50 13 burnette et al. (2020) 16 ehrlinger, j., plant, e. a., hartwig, m. k., vossen, j. j., columb, c. j., & brewer, l. e. do gender differences in perceived prototypical computer scientists and engineers contribute to gender gaps in computer science and engineering? 2018 51 12.75 10 ehrlinger et al. (2018) 17 bers, m. u. coding as another language: a pedagogical approach for teaching computer science in early childhood 2019 50 16.67 50 bers (2019) 18 malik, s. i., & al-emran, m. social factors influence on career choices for female computer 2018 49 12.25 25 malik & al-emran (2018) al husaeni et al. mapping visualization analysis of computer science… | 8 no authors title year cites cites per year cites per author refs science students. 19 nissim, k., bembenek, a., wood, a., bun, m., gaboardi, m., gasser, u., o'brien, d.r., steinke, t. & vadhan, s. bridging the gap between computer science and legal approaches to privacy 2017 48 9.60 8 nissim et al. (2017) 20 hur, j. w., andrzejewski , c. e., & marghitu, d. girls and computer science: experiences, perceptions, and career aspirations 2017 48 9.60 16 hur et al. (2017) 3.2. research developments in computer science research the development of research on computer science over the last 5 years, namely from 2017-2021 which has been published in google scholar indexed publications amounted to 992 articles. the number of each publication in sequence from 2017 to 2021 is 198, 232, 208, 206, and 148 articles. table 2 also shows that the most researched and published articles on computer science in 2018 amounted to 232 articles and the least research occurred in 2021, namely 148 articles. the average publication for the last 5 years is 198.4 articles. the development of research on computer science is shown more clearly in fig. 1. table 2. development of computer science research. year number of publication per year 2017 198.0 2018 232.0 2019 208.0 2020 206.0 2021 148.0 total 992.0 averange 198.4 fig. 1 shows that in 2017 research on computer science there were 198 articles and there was an increase in publications in 2018 to 232 articles. however, it 9 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 decreased in 2019 to 208 articles. research on computer science continues to decline from 2018, namely 2020, there were 206 articles and 148 articles in 2021. overall, it can be seen that the increase occurred only in 2018 only. in the following year, it continued to decline. fig. 1. level of development in computer science research. 3.3. mapping analysis based on text data of computer science using vosviewer in mapping the analysis based on text data using the vosviewer application. found 5792 terms relevant to computational thinking research. however, in this study, we give the minimum number of occurrences of the term to be 10 times. therefore, the results obtained are 136 items used in the process of mapping data analysis. research related to computer science based on network visualization is divided into 5 clusters and there are 5028 links. cluster 1 has 37 items, marked in red and presented in fig. 2. the items in cluster 1 are access, assessment, challenge, classroom, computational thinking, computer science curriculum, computer science education, computer science educator, computer science program, computer science teacher, content, curricula, curriculum, development, effort, equity, evaluation, faculty, framework, implementation, implication, information technology, knowledge, learner, learning, opportunity, perception, practice, program, project, school survey, teacher, teaching, technique, tool, and understanding. cluster 2 has 36 items and is marked in green, shown in fig. 3. the items in cluster 2 are algorithm, application, approach, area, artificial intelligence, aspect, computer, computer science, computer science department, computer science perspective, computer science research, computer scientist, computing, data, data science, discipline, fact, field, focus, mathematics, model, paper, perspective, physics, problem, process, research, researcher, science, social science, system, technology, term, theoretical computer science, theory, and topic. al husaeni et al. mapping visualization analysis of computer science… | 10 fig. 2. network visualization of the main term in cluster 1. fig. 3. network visualization of the main term in cluster 2. cluster 3 has 32 items and is marked in blue, shown in fig. 4. the items in cluster 3 are case, case study, change, college, computer engineering, computer science concept, computer science course, computer science degree, a computer science major, computer science student, course, department, effectiveness, engineering experience, factor, high school, higher education, idea, impact, introductory computer science, introductory computer science course, 11 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 investigation, software, software engineering, strategy, student, study, success, systematic literature review, time, and university. cluster 4 has 17 items marked in yellow, shown in fig. 5. the items in cluster 4 are analysis, attention, context, demand, education, effect, gender, group, importance, industry, initiative, level, participation, question, relationship, and role. fig. 4. network visualization of the main term in cluster 3. fig. 5. network visualization of the main term in cluster 4. cluster 5 has 14 items and is marked in purple (fig. 6). the items in cluster 5 are ability, activity, attitude, career, computer science class, concept, goal, information, motivation, programming, review, skill, stem, and subject. al husaeni et al. mapping visualization analysis of computer science… | 12 fig. 6. network visualization of the main term in cluster 5. in mapping analysis using vosviewer, cluster describes the relationship between one term and another (nandiyanto et al., 2021; al husaeni & nandiyanto, 2022; nandiyanto & al husaeni, 2021). the existing terms are labeled and also different colors. the color indicates the term cluster is located. the size of each label indicates the frequency with which the term appears or is used in computer science research. if the size of the circle label is bigger, the more often the term is used, and vice versa, the smaller it is, the less often it is used (nandiyanto et al., 2021; al husaeni & nandiyanto, 2022; nandiyanto & al husaeni, 2021). fig. 7 illustrates the network visualization in mapping analysis with vosviewer. network visualization shows the relationship from one term to another and shows the occurrences of that term. based on fig. 7, it can be seen that the term computer science has the largest label size. this shows that the term computer science has a high frequency of occurrences and the connection with other terms is also high. 13 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 fig. 7. network visualization of computer science research. in this study, we found 4 main terms that have a fairly high degree of connectedness with computer science terms, namely computer science education term with link strength of 116 (fig. 8a), student term with link strength of 221 (fig. 8b), education term with link strength of 78 (fig. 8c), and skill term with link strength 58 (fig. 8d). the link strength range of terms that are related to research in the field of computer science can be seen in fig. 9. fig. 9 shows that research with the theme of computer science has the highest correlation with the student term. this shows that many researchers are researching computer science and it is related to the student condition or term student. al husaeni et al. mapping visualization analysis of computer science… | 14 fig. 8. network the relationship between computer science research and other terms (a) computer science to computer science education; (b) computer science to students; (c) computer science to education; and (d) computer science to skills. fig. 9. the link strength range of terms that are related to research in the field of computer science. fig. 10 illustrates the overlay visualization of research in the field of computer science from 2017 to 2021. the overlay visualization shows the novelty of research on related terms (nandiyanto et al., 2021; al husaeni & nandiyanto, 2022; nandiyanto & al husaeni, 2021). many types of research on computer science have been carried out in the 20182019 timeframe as shown in fig. 11. the 15 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 term computer science has the largest research time in 2018.8-2019.0. therefore, when computer science research is in 2020-2021 there are still many opportunities to get new research. mapping analysis on overlay visualization data using vosviewer can be used as a reference for new research with the theme of computer science (nandiyanto et al., 2021; al husaeni & nandiyanto, 2022; nandiyanto & al husaeni, 2021). fig. 10. overlay visualization of computer science research on 2017-2021. fig. 11. overlay visualization of computer science research on 2018-2019. al husaeni et al. mapping visualization analysis of computer science… | 16 3.4. mapping analysis based on bibliographic data of computer science using vosviewer mapping analysis based on bibliographic data was conducted to see which authors contributed the most to the field of computer science research published and indexed by google scholar. fig. 12 shows the network visualization author with the most contributions to the collected data. the data shows that goode, j has the most contribution to research in the field of computer science in 2017-2021 which is published and indexed by google scholar, which contributes 12 published article documents. fig. 12. network visualization of the author in computer science research. from the results of this research, we can look for several topics of research in computer science education and their relationship to several other fields of discussion. we can also determine research themes that are more recent and in accordance with research trends in the year concerned, by looking at the track record of previous research. 4. conclusion the publish or perish 7 application was used to collect data from the google scholar database for the period 20172021. the keyword "computer science" was used to collect data. the data search yielded 992 articles that were thought to be relevant. with a total of 232 papers, the results showed that computer science research was quite popular in 2018. in the years 2019-2021, there was a decrease in computer science research. computer science terms are linked to four key terms in each cluster, according to the mapping analysis performed with the vosviewer application: student, computer science 17 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 education, education, and skills. the term "student" is commonly connected with computer science research, specifically the strength of link 221. references al husaeni, d. f., & nandiyanto, a. b. d. (2022). bibliometric using vosviewer with publish or perish (using google scholar data): from step-by-step processing for users to the practical examples in the analysis of digital learning articles in pre and post covid-19 pandemic. asean journal of science and engineering, 2(1), 1946. alhazov, a., boian, e., burtseva, l., ciubotaru, c., cojocaru, s., colesnicov, a., demidova, v., ivanov, s., macari, v., magariu, g. & malahov, l. (2010). investigations on natural computing in the institute of mathematics and computer science. computer science journal of moldova, 53(2), 101-138. aria, m., & cuccurullo, c. (2017). bibliometric: an r-tool for comprehensive science mapping analysis. journal of informetrics, 11(4), 959-975. armoni, m., & gal-ezer, j. (2014). early computing education: why? what? when? who?. acm inroads, 5(4), 54-59. bers, m. u. (2019). coding as another language: a pedagogical approach for teaching computer science in early childhood. journal of computers in education, 6(4), 499528. bonham, k. s., & stefan, m. i. (2017). women are underrepresented in computational biology: an analysis of the scholarly literature in biology, computer science and computational biology. plos computational biology, 13(10), e1005134. borrego, c., fernández, c., blanes, i., & robles, s. (2017). room escape at class: escape games activities to facilitate the motivation and learning in computer science. jotse, 7(2), 162-171. burnette, j. l., hoyt, c. l., russell, v. m., lawson, b., dweck, c. s., & finkel, e. (2020). a growth mind-set intervention improves interest but not academic performance in the field of computer science. social psychological and personality science, 11(1), 107-116. cobo, m. j., lópez‐herrera, a. g., herrera‐viedma, e., & herrera, f. (2012). scimat: a new science mapping analysis software tool. journal of the american society for information science and technology, 63(8), 1609-1630. ehrlinger, j., plant, e. a., hartwig, m. k., vossen, j. j., columb, c. j., & brewer, l. e. (2018). do gender differences in perceived prototypical computer scientists and engineers contribute to gender gaps in computer science and engineering?. sex roles, 78(1), 40-51. fields, d. a., kafai, y., nakajima, t., goode, j., & margolis, j. (2018). putting making into high school computer science classrooms: promoting equity in teaching and learning with electronic textiles in exploring computer science. equity and excellence in education, 51(1), 21-35. al husaeni et al. mapping visualization analysis of computer science… | 18 garcia, r., falkner, k., & vivian, r. (2018). systematic literature review: self-regulated learning strategies using e-learning tools for computer science. computers and education, 123, 150-163. giannakos, m. n., pappas, i. o., jaccheri, l., & sampson, d. g. (2017). understanding student retention in computer science education: the role of environment, gains, barriers and usefulness. education and information technologies, 22(5), 2365-2382. hill, j. t., demarest, b. l., bisgrove, b. w., gorsi, b., su, y. c., & yost, h. j. (2013). mmappr: mutation mapping analysis pipeline for pooled rna-seq. genome research, 23(4), 687-697. hur, j. w., andrzejewski, c. e., & marghitu, d. (2017). girls and computer science: experiences, perceptions, and career aspirations. computer science education, 27(2), 100-120. jurado de los santos, p., moreno-guerrero, a. j., marín-marín, j. a., & soler costa, r. (2020). the term equity in education: a literature review with scientific mapping in web of science. international journal of environmental research and public health, 17(10), 3526. leyton-brown, k., milgrom, p., & segal, i. (2017). economics and computer science of a radio spectrum reallocation. proceedings of the national academy of sciences, 114(28), 7202-7209. malik, s. i., & al-emran, m. (2018). social factors influence on career choices for female computer science students. international journal of emerging technologies in learning, 13(5), 56-70 mennis, j. l., & jordan, l. (2005). the distribution of environmental equity: exploring spatial nonstationarity in multivariate models of air toxic releases. annals of the association of american geographers, 95(2), 249-268. nandiyanto, a. b. d., & al husaeni, d. f. (2021). a bibliometric analysis of materials research in indonesian journal using vosviewer. journal of engineering research. 9(asseee special issue), 1-16. nandiyanto, a. b. d., al husaeni, d. n., & al husaeni, d. f. (2021). a bibliometric analysis of chemical engineering research using vosviewer and its correlation with covid-19 pandemic condition. journal of engineering science and technology, 16(6), 4414-4422. nissim, k., bembenek, a., wood, a., bun, m., gaboardi, m., gasser, u., o'brien, d.r., steinke, t. & vadhan, s. (2017). bridging the gap between computer science and legal approaches to privacy. harvard journal of law & technology, 31, 687. passey, d. (2017). computer science (cs) in the compulsory education curriculum: implications for future research. education and information technologies, 22(2), 421-443. qian, y., hambrusch, s., yadav, a., & gretter, s. (2018). who needs what: recommendations for designing effective online professional development for 19 | international journal of informatics information system and computer engineering 3(1) (2022) 1-18 computer science teachers. journal of research on technology in education, 50(2), 164-181. sax, l. j., lehman, k. j., jacobs, j. a., kanny, m. a., lim, g., monje-paulson, l., & zimmerman, h. b. (2017). anatomy of an enduring gender gap: the evolution of women’s participation in computer science. the journal of higher education, 88(2), 258-293. talen, e. (1998). visualizing fairness: equity maps for planners. journal of the american planning association, 64(1), 22-38. tussyadiah, i. (2020). a review of research into automation in tourism: launching the annals of tourism research curated collection on artificial intelligence and robotics in tourism. annals of tourism research, 81, 102883. vakil, s. (2018). ethics, identity, and political vision: toward a justice-centered approach to equity in computer science education. harvard educational review, 88(1), 26-52. wang, d., liang, y., xu, d., feng, x., & guan, r. (2018). a content-based recommender system for computer science publications. knowledge-based systems, 157, 1-9. webb, m., davis, n., bell, t., katz, y. j., reynolds, n., chambers, d. p., & sysło, m. m. (2017). computer science in k-12 school curricula of the 2lst century: why, what and when?. education and information technologies, 22(2), 445-468. weintrop, d. (2019). block-based programming in computer science education. communications of the acm, 62(8), 22-25. weintrop, d., & wilensky, u. (2017). comparing block-based and text-based programming in high school computer science classrooms. acm transactions on computing education (toce), 18(1), 1-25. wolch, j., wilson, j. p., & fehrenbach, j. (2005). parks and park funding in los angeles: an equity-mapping analysis. urban geography, 26(1), 4-35. xiang, h., lee, c., koo, h. j., gong, x., & whangbo, m. h. (2013). magnetic properties and energy-mapping analysis. dalton transactions, 42(4), 823-853. sathish kumar et al. enhanced wearable strap for feminine using iot | 80 enhanced wearable strap for feminine using iot sathish kumar, s. nandhini, r. sujitha department of computer science and engineering, manakula vinayagar institute of technology, india *corresponding email: sathishmail8@gmail.com a b s t r a c t s a r t i c l e i n f o women are increasingly experiencing a slew of security difficulties while traveling alone at night, particularly in the it industry. despite the benefits of new technologies, the rate of crime against women continues to rise. even though numerous security gadgets are available on the market, women are unaware of them and do not use them. we are going to develop a prototype design using iot. the term "physical objects” that are outfitted with sensors, computing power, software, and other technologies that enable them to connect and exchange data with other devices and systems over the internet or other communication networks are referred to as the internet of things. our project's goal is to give protection to working women and children. most crimes occur as a result of a person's lack of awareness. we plan to keep the person aware throughout by administering a vibration. the device comprises components such as the start button, arduino uno, panic button, gsm neo 6 m, gps, pulse sensor, vibration motor, 6v transformer, and touch sensor. the current location is determined using gprs and gsm neo 6 m. in emergency cases, a pulse sensor detects the person's actual pulse rate. the person must use a touch sensor to stop the vibration. if the vibration does not stop within the specified time, it is assumed that the person is not in an active state. the emergency alert is then sent to the predefined contacts stored on the arduino board. the transformer with a voltage range of article history: received 25 may 2022 revised 30 may 2022 accepted 10 june 2022 available online 26 june 2022 aug 2018 __________________ keywords: arduino uno, panic button, gps neo-6m, gsm sim800c, vibration motor, pulse sensor international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 80-92 81 | international journal of informatics information system and computer engineering 3(1) (2022) 80-92 6v supplies power to the entire device. the future scope of this project is we can collect datasets of all hospitals for emergency purpose, creating offline maps to locate the victim without internet connection. then the device can also contain mic and camera to live monitor the consequences. 1. introduction in terms of women's security, we are most likely living in the darkest period in our contemporary society's history. we want to make women feel as powerful as they have always been, strong enough to combat the parasites of our society, strong enough to overcome obstacles, and strong enough to protect themselves from sexual attacks. we want to empower those who are essential to our survival. our goal is to create a system that will make every location and hour safer for women once again. a mechanism that will restore humanity's communal nature. with a single touch of a button, this technology will geotag and send an sos alert to the local police station, close contacts, and anyone in and around the crime scene. the goal is to compensate for the time it takes the cops to get at the scene. most of the incidents were happened in public places especially at train station, market, public gathering etc. the below fig.1 explains the pie chart which implies the percentage of incidents happened in public places. fig. 1. percentage of incidents happened in public places the percentage of cases happened in different countries where thailand has the highest percentage that women has suffered physical assault compared to other countries like united kingdom, brazil, india. the fig.2 bar chart represents the percentage of physical assault faced by women in different countries fig. 2. percentage of women suffered physical assaults according to the survey the percentage of cases increases in last five years especially women between the age 18-24 faced many consequences like harassment, so sathish kumar et al. enhanced wearable strap for feminine using iot | 82 the below fig.3 shows the bar chart with percentage of cases and different ages of women who faced the harassment. fig. 3. most 18 – 24 years women has been harassed in public places in last five years 2. literature survey ”smart wearable device for women safety using iot”: v. hyndavi, n. sai nikhita, s. rakesh – 2021. any woman must touch that push button if she feels insecure or in danger. the victim's image will then be simultaneously taken and saved on this raspberry pi operating system (os). then, the previously saved os image will be taken out automatically because adding more photographs can increase storage (hyndavi et al., 2020). “design and development of an advanced affordable wearable safety device for women”: freedom against fearsome”: israt humaira, kazi arman ahmed, sayantee roy, zareen tasnim safa, f. m. tanvir hasan raian , md. ashrafuzzaman – 2021. this device is simple in design, affordable, and practicable. by providing a hidden camera, gps and gsm module, shocking device, and voice recognizer, our products "bohnni" and "badhon" will ensure a safe atmosphere for women and children in any scenarios. there are already a variety of health and activity monitoring sensor belts on the market, but the design of an integrated safety device is fundamentally different since very accurate calculations and designs are required (humaira et al., 2021). ”iot based wearable device to monitor the signs of quarantined remote patients of covid-19”: nizar al bassam, shaik asif hussain, ammar al qaraghuli ,jibreal khan , e.p. sumesh, vidhya lavanya – 2021.internet of things (iot) technology have enabled remote health monitoring systems to use, accessible, and handy for measuring and recording patient parameters in a comfortable setting. the proposed iot-based health monitoring system can measure physiological parameters and health symptoms in covid-19-affected patients and transmitting those data to an application peripheral interface (api), which serves as a database for browsing and monitoring the infection level (bassam et al., 2021). ”women self defense device”: jayashree agarkhed, aishwarya rathi, maheshwari, faqarunnisa begum – 2021. our main intention in the current circumstance is: "women should always feel free to roam; a camera and a voice recognition module may also be installed in a watch so that it catches and records all of the photos and voice said by the victim, which are then delivered to the registered numbers and saved immediately (agarkhed et al., 2020). “smart security device for women based on iot using raspberry pi” prottasha ghosh, tanjimmasroor bhuiyan, muhib ashraf nibir , md. emran hasan , md. rabiul islam , md. rokib hasan , tanvir hossain – 2021. this gadget may be changed with a clever and effective one, 83 | international journal of informatics information system and computer engineering 3(1) (2022) 80-92 as security is a big issue for any woman in today's environment. for the first time, this sort of gadget has been used to capture direct evidence and store it on the web for later use (ghosh et al., 2021). ”iot based unified approach for women safety alert using gsm”: venkatesh. k, parthiban. s, santhosh kumar. p, vinoth kumar.c.n. s: 2021. the goal of this project is to ensure that every lady in our general population believes that everything in the world is ok and secure. according to a survey in india, 53 out of 100 working women do not believe that everything is ok in the world women are working at night. women experiencing barriers are also prevalent in delhi, mumbai, hyderabad,kolkata, and pune, according to a survey of 86 working women in india (venkatesh et al., 2021). “arduino based smart shoe system for women safety, defense and integrated intelligent tracking”: syed mosharaf hossain, subhendu chakraborty – 2021. this smart safety shoe for ladies is a ready-to-wear gadget for women's everyday use. to the best of the writers' knowledge, there is no such off-the-shelf, ready-to use equipment that can be utilised by women for safety purposes. there is no need for a complicated charging circuit with this gadget, and no wires can be seen from the exterior. the main purpose of this project was to create a ready-to-use and portable smart ladies’ safety gadget (hossain et al., 2021). “iot based women securitya contemplation”:deepinder kaur, ravitachahar, jatinder ashta – 2020. in comparison to mobile applications, preprogrammed smart hardware devices can perform efficiently in danger circumstances. in the future, a new device can be designed for women who use public transportation in which the female and driver must give biometric impressions before beginning the trip and information of both is used by gps in case of an unfortunate event, regardless of whether the passenger has a smart phone or not (kaur et al., 2020). “smart safety and security solution for women using knn algorithm and iot”:bysani sai yaswanth, darshan r s, pavan h, srinivasa d b, b t venkatesh murthy – 2020. this study focuses on an iot-based self-security system that is comfortable, easy to use, and wearable, and that assists in sharing the user's position when they are panicked,as well as locating the closest safe area. the user's position is sent into the machine learning system, which then forecasts the location of the nearest safe spot (yaswanth et al., 2020). “lifecraft: anandroid based application system for women safety”:rabbina ridankhandoker, shahreen khondaker, fatiha-tus-sazia, fernaznarin nur, shaheena sultana – 2019.this study presents a new model of women's safety that attempts to create a particularly safe environment for women. this study compares the essential requirements of an intelligent security system with technological demand and system construction problems. it may beemployed in a variety of situations, including sexual problems, accidents, hijackings, and public attacks (khandoker et al., 2019). “protecht – implementation of an iot based 3 –way women safety device”: trisha sen, arpita dutta, shubham singh,vaegaenveen kumar – 2019. the iot-based 3-way women's safety device accomplishes its goal by offering users sathish kumar et al. enhanced wearable strap for feminine using iot | 84 with a self-defense mechanism as well as facilities for collecting information and recording evidence. when the user says the words "emergency" in the android app, the savedemergency contacts receive a message and a call, as well as the user's current position. the gadget responds quickly and may assist the user in remaining safe in any situation (sen et al., 2019). “smarisa: a raspberry pi based smart ring for women safety using iot”navya r sogi, priya chatterjee, nethra u, suma v – 2018. this report proposed a method through which a woman can immediately notify the appropriate authorities if she is in danger. the suggested method obtains the device's coordinates by using gps tracking of the smart phone. this method also uses the image's url and an alert message to notify the family and police officers (sogi et al., 2018). ”b with uiot based woman security system”: akhila c m, jisna raju, lakshmi sethunath, vishnu prasad yadav, mrs. manju bravada’s – 2019. we have now created a project that will address crucial concerns that women confront in today's society and give them the strength to fight back. it includes a band prototype, shock shoe, and mobile device that displays the current location on a google map as well as live streaming on the internet (akhila et al., 2019). “design of a smart safety device for women using iot”: wasim akram, mohit jain, c. sweetlinhemalatha 2019. this proposes a solution that will attempt to address the shortcomings of the current system while also ensuring women's false proof safety. safe spots from the victim's present position will be shown on the map so that women may reach the safe place from their current location, according to a smartphone app created for women's protection (akram et al., 2019). “the personal stuna smart device for women’s safety”: shivani ahir, smit kapadia, prof. jigar chauhan, prof. nidhi sanghavi – 2018. they concentrated on creating a smart bracelet that can be triggered by tapping the screen twice. the pulse sensor is an integrated sensor in the gadget that detects the person's pulse rate. the gadget now has a force sensor that activates when it is thrown with force. on the android platform, a smart application will be built that connects to the device through bluetooth interface and displays the subject's detected data to the ice contacts (ahir et al., 2018). “safeband: a wearable device for the safety of women in bangladesh”: muhammad nazrul islam, nuzhat tabassum promi, jannatul maowa shaila, mohoshina akter toma, maria afnan pushpo, fatema binte alam ,syeda nusraht khaledur, tasmiah tamzid anannya, md. fazle rabbi – 2018. the 'safe band' system's assessment research revealed that the system efficiently executed all the capabilities and had a very good usability quality. in the future, it is planned to merge the entire gadget onto a single mini chip for real-world application, making transporting the system more efficient for women (islam et al., 2018). ”design and development of an iot based wearable device for the safety and security of women and girl children”: anandjatti, madhvikannan, alisha rm, vijayalakshmi p, shresthasinha 2018. 85 | international journal of informatics information system and computer engineering 3(1) (2022) 80-92 they focused on using the internet of things in conjunction with wearable devices to ensure the protection and security of women and girls. body position is estimated using raw accelerometer data from a triple axis accelerometer (jatti et al., 2016). 3. objective of the research women are not safe anywhere and are most vulnerable when traveling alone into lonely roads and deserted places, so our purpose of effort is to reduce or overcome the rate of crime against women. the development of smart band for women safety is to keep the victim safe from the danger. when a woman or a child feels threatened, she must activate the device by pressing the start button. after activating the device, press the panic button to send the message "help" along with the current location to the predefined contacts via gsm neo 6m and gprs. at the same time, the pulse sensor detects the person's pulse rate. when the live location is sent, the person's pulse rate is also sent. in addition to this we have placed vibration motor to always keep the person alert. the vibration motor is placed to always keep the person alert. when in the active state, the person must use the reset button to stop the vibration. if the vibration motor is not stopped within a certain period, assuming the person is not conscious, an emergency alert is sent to the predefined contacts with the subject "emergency". the device receives power from the transformer ranges of 6v. 4. existing work when women are traveling alone or in a distant region where they cannot acquire help or appropriate assistance, they are attacked (sogi et al., 2018). this research proposes an internet of things-based solution to the problem of women's safety that overcomes present technological restrictions. when a woman is in danger, the proposed design incorporates mechanisms that warn family members and a nearby police station so that help may be sent quickly. a shock waves generator is also included in the suggested design, which women may use to assault the perpetrator. some of the other components of the projected work that give further assistance to women are as follows: 1. sending group sms with the device and the victims' phones. 2. the victim's voice is recorded and later used as evidence against the offender. 3. locating a safe spot on a map from where the victim is now. 5. design of the existing work the suggested women's safety device supports a woman who is in a potentially perilous situation. the device is practically ready for everything that might go against the woman's wishes. an at mega 328 microcontroller oversees it. the gadget is activated by a fingerprint scanner, a gsm (global system for mobile communications) module transmits alarm messages, a buzzer warns the environment, and a shock wave generator is utilized for selfdefence. on an lcd, the message is shown. the following fig.4 explains how the gadget works: the woman's fingerprint must first be registered before the device can be activated. the gadget continues scanning the woman's fingerprint every minute as soon as she sathish kumar et al. enhanced wearable strap for feminine using iot | 86 turns it on. the gadget will be activated if the scanner finds no fingerprint, blowing a bell to alert the public nearby. because it only scans the woman's fingerprint in an emergency, i.e., when she detects a threat, the device's quality is unaffected. the gps data is also relayed to the lcd and the gsm modem, which will deliver the message to the woman's relatives and friends (sogi et al., 2018). even if she is pulled down from behind and is unable to activate an alarm, the device will send an emergency message to all the woman's ice contacts (in case of emergency contacts) alerting them to her current position. the design also includes a shock wave generator that may be used as a weapon or to assist the lady in self-defence. a shock wave generator's gear, which comprises a switch, transformer, and wires. one of the two free ends of the wires acts as the high voltage source, while the other serves as the return path's ground. the high voltage cannot arc unless it comes into touch with the attacker's body, which acts as a conducting conduit between the two ends because the loose ends are not in direct contact. the circuit has three important stages. •the oscillator •the voltage amplifier •the power supply when the battery is fully charged, the voltage is delivered to the oscillator stage. the transformer serves as an inverter, increasing the frequency of oscillation. the output of the transformer is then delivered to the capacitors, which hold the current until it is utilized to electrocute the aggressor. in the existing work, they have added some features which supports through android application, and it is represented in the below fig.5. fig. 4. workflow of the proposed design fig. 5. additional features supported through android application 6. proposed work our suggested smart band project consists of hardware components connected to an arduino uno. when the panic button is pressed, the gps and gsm module activates, tracking the user's current location and sending a message with the victim's heart rate via the pulse sensor. our primary goal is to keep the victim awake. to do this, we provide a vibration motor for a set period. once the victim awakens, they can manually switch off the vibration, and if it continues, another alarm 87 | international journal of informatics information system and computer engineering 3(1) (2022) 80-92 message is sent to the predetermined contact. a switch is used to turn on the device. a push button is triggered to track the location and send message and further processing (see figs. 6-10). module 1: activating the device and tracing the location using gps neo-6m module. module 2: sending messages to the predefined contact. module 3: pulse rate detection of the person. module 4: finding the active state of the person using vibration motor. a) proposed architecture fig. 6. architecture diagram b.) module 1: activating the device and tracing the location using gps neo-6m module global positioning system neo-6m: parts used: 1. arduino uno board 2. neo 6m gps module 3. jumper wire connections: 1. first, the rx pin from the gps module needs to be connected to arduino pin 8. 2. the gps module's tx pin should then be linked to pin 9 on the arduino. 3. the gps module's vcc pin should then be linked to the 12v power supply. 4. •the gps's gnd is then linked to the power supply's gnd. fig. 7. diagrammatic view of tracking and sending location c.) module2: sending messages to the predefined contacts global system for mobile communication sim 800c: parts used: 1. arduino uno board 2. sim 3. jumper wires connections: 1. •the sim card must first be inserted into the sim 800c in the sim cardholder. sathish kumar et al. enhanced wearable strap for feminine using iot | 88 2. •the arduino pin 4 should be linked to the p in the gsm. 3. •the arduino pin 3 should be linked to the t in the gsm. the panic button was connected to the vcc. and the arduino gnd was linked by the gnd. d). working discription: according to the above fig.7, it explains how the gps and gsm connected with the arduino. the start switch must first be used to turn on the gadget. the panic button must then be held down for 5 seconds. following the gps tracking of the current location, the tracked location is communicated via the gsm module 800c to the designated contacts along with the message "help me." e). module 3: pulse rate detection of the person the arduino uno board has a pulse sensor attached to it. parts used: 1. arduino uno board 2. pulse rate sensor 3. jumper wires. connections: 1. the pulse sensor's a0 is linked to the arduino uno's a0. 2. the arduino uno's gnd is connected to the pulse sensor's gnd. 3. the pulse sensor's vcc is connected to the start button fig.8 code of gsm sending sms f). workings discription: according to the above fig.8, it explains how gsm is connected to the arduino and it shows the workflow of how gsm sends messages to the predefined contact. the person's pulse rate is frequently monitored by the pulse sensor. the person's pulse rate is communicated to the designated contacts as soon as the panic button is pressed, coupled with the person's location. the alert message "emergency need help" is delivered to the predefined contacts once more if the pulse rate changes to the left or right of the threshold level. g). module 4: finding the active state of the person using vibration motor parts used: 1. vibration motor 2. arduino uno board 3. jumper wires connections: 1. the arduino uno's gnd and the vibration motor's gnd are connected. 2. the arduino pin 7 is connected to the vcc of the vibration motor.11 fig.9 code of vibration motor h). working discription: according to the above fig.9, it explains how vibration motor is connected with the arduino and it represents how the vibration will function. the vibration motor is preserved to detect a person's active state. in a gap of 30 minutes, it vibrates for 5 seconds. if the person is in an active state, she must press the button to cease the vibration. say an alert message is sent to the predetermined contacts if the vibration hasn't stopped for a while. 89 | international journal of informatics information system and computer engineering 3(1) (2022) 80-92 i). our prototype fig. 8. our actual prototype j). message along with the bpm value fig. 9. gps location gps location fig. 10. gps location 7. result and discussion a) device price comparison we have taken the survey based on price comparison from the existing device and it is described in the below fig.11 fig. 11. based on cost comparison b ) pulse rate detection a person pulse rate is detected with the help of pulse sensor. the below bar chart in fig.12 analysis the pulse rate of the person based on their ages and their average bpm is shown. fig. 12. pulse rate detection level c) vibration analysis sathish kumar et al. enhanced wearable strap for feminine using iot | 90 the rate of vibration is calculated with the peak values. it represents the peak-topeak values in the form of waves, according to the period which is shown in the fig. 13. fig. 13 vibration rate level the actual rate of vibration which was detected by our device is represented in the below fig. 14. fig. 14. vibration detection rate 8. conclusion our main intention in the current era is that “women should always feel free to roam; wherever; anytime. they should not experience any adverse circumstances of any type. as a result, we created the smart band, which consists of several hardware parts connected to iot devices. since the device features a gps and gsm module, it can track where abouts and send messages to contacts that have been predefined. a vibration motor is also utilized in addition to keeping the sufferer awake for a set amount of time. we believed that the women would benefit greatly from being in difficult situations. this device's future applications will benefit youngsters, physically challenged ladies, and senior citizens. this project can be utilized by parents to find where their children are in addition to enhancing the conditions for women's safety in india. additionally, we may improve by compiling hospital datasets that will aid the sufferer by using gps and gsm to follow her location and send signals to the closest hospital when her pulse drops, or she needs an emergency care. we may also create offline maps to locate the victim even when there is no signal or internet connection. we can also include camera and microphone in addition to live monitor the activities when need. references agarkhed, j., rathi, a., & begum, f. (2020, october). women self defense device. in 2020 ieee bangalore humanitarian technology conference (b-htc) (pp. 1-5). ieee. ahir, s., kapadia, s., chauhan, j., & sanghavi, n. (2018, january). the personal stuna smart device for women's safety. in 2018 international conference on smart city and emerging technology (icscet) (pp. 1-3). ieee. akhila, c. m., raju, j., sethunath, l., yadav, v. p., & bhavadas, m. m. (2019). b with u-iot based woman security system. vol, 8, 507-509. 91 | international journal of informatics information system and computer engineering 3(1) (2022) 80-92 akram, w., jain, m., & hemalatha, c. s. (2019). design of a smart safety device for women using iot. procedia computer science, 165, 656-662. al bassam, n., hussain, s. a., al qaraghuli, a., khan, j., sumesh, e. p., & lavanya, v. (2021). iot based wearable device to monitor the signs of quarantined remote patients of covid-19. informatics in medicine unlocked, 24, 100588. ghosh, p., bhuiyan, t. m., nibir, m. a., hasan, m. e., islam, m. r., hasan, m. r., & hossain, t. (2021, january). smart security device for women based on iot using raspberry pi. in 2021 2nd international conference on robotics, electrical and signal processing techniques (icrest) (pp. 57-60). ieee. hossain, s. m., & chakraborty, s. (2021). arduino based smart shoe system for women safety, defense and integrated intelligent tracking, www.ijaceeonline.com. 6(1), pp. 1–7. humaira, i., ahmed, k. a., roy, s., safa, z. t., raian, f. m. t. h., & ashrafuzzaman, m. (2021). design and development of an advanced affordable wearable safety device for women: freedom against fearsome. adv. sci., technol. eng. syst. j., 6(2), 829-836. hyndavi, v., nikhita, n. s., & rakesh, s. (2020, june). smart wearable device for women safety using iot. in 2020 5th international conference on communication and electronics systems (icces) (pp. 459-463). ieee. islam, m. n., promi, n. t., shaila, j. m., toma, m. a., pushpo, m. a., alam, f. b., ... & rabbi, m. f. (2018, november). safeband: a wearable device for the safety of women in bangladesh. in proceedings of the 16th international conference on advances in mobile computing and multimedia (pp. 76-83). jatti, a., kannan, m., alisha, r. m., vijayalakshmi, p., & sinha, s. (2016, may). design and development of an iot based wearable device for the safety and security of women and girl children. in 2016 ieee international conference on recent trends in electronics, information & communication technology (rteict) (pp. 1108-1112). ieee. kaur, d., chahar, r., & ashta, j. (2020, march). iot based women security: a contemplation. in 2020 international conference on emerging smart computing and informatics (esci) (pp. 257-262). ieee. khandoker, r. r., khondaker, s., nur, f. n., & sultana, s. (2019, december). lifecraft: an android based application system for women safety. in 2019 international conference on sustainable technologies for industry 4.0 (sti) (pp. 1-6). ieee. sen, t., dutta, a., singh, s., & kumar, v. n. (2019, june). protecht–implementation of an iot based 3–way women safety device. in 2019 3rd international conference on electronics, communication and aerospace technology (iceca) (pp. 1377-1384). ieee. sogi, n. r., chatterjee, p., nethra, u., & suma, v. (2018, july). smarisa: a raspberry pi based smart ring for women safety using iot. in 2018 international conference on inventive research in computing applications (icirca) (pp. 451454). ieee. http://www.ijaceeonline.com/ sathish kumar et al. enhanced wearable strap for feminine using iot | 92 venkatesh, k., parthiban, s., kumar, p. s., & kumar, c. v. (2021, february). iot based unified approach for women safety alert using gsm. in 2021 third international conference on intelligent communication technologies and virtual mobile networks (icicv) (pp. 388-392). ieee. yaswanth, b. s., darshan, r. s., pavan, h., srinivasa, d. b., & murthy, b. v. (2020, december). smart safety and security solution for women using knn algorithm and iot. in 2020 third international conference on multimedia processing, communication & information technology (mpcit) (pp. 87-92). ieee. 11 | international journal of informatics information system and computer engineering 4(1) (2023) 11-22 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 computational thinking: the essential skill for being successful in knowledge science research adam mukharil bachtiar school of knowledge science, japan advanced institute of science and technology, 1-8 asahidai, nomi shi, ishikawa, japan *corresponding email: s2220401@jaist.ac.jp 1. introduction we are living in society 5.0 with a focus on increasing the capability of human-being to create innovation (carayannis & morawska-jancelewicz, 2022). the limitation between the real and digital world seems to fade out. the interaction among humans is very intense in social media. digitalization is a word that the constituent parties in society should fulfill. higher education plays a significant role in ensuring the availability of human resources with requisite skills (shin. j.c., 2015). digitalization leads society to a new concept called the vuca world (mack & a b s t r a c t s a r t i c l e i n f o the vuca world concept was established in 2016 as the new challenge universe in the 21st century. humans live in society 5.0 and the vuca world simultaneously. the digital word has been a noisy word since then. there are a lot of requisite skills to be a survival kit for this kind of era. the vuca world's affection is spreading in the way of thinking and creating innovation, especially in the research domain. as a newcomer, knowledge science should state the requisite skills for its researchers to conduct their research successfully. many researchers offered computational thinking as a candidate for an essential skill to satisfy the effect of the vuca world. this study was focused on conducting a descriptive analysis method based on several literature reviews for mapping how computational thinking can serve as a best practice for knowledge science research. this study successfully revealed the connection between computational thinking. article history: submitted/received 21 jan 2023 first revised 23 feb 2023 accepted 27 march 2023 first available online 09 april 2023 publication date 01 june 2023 aug 2018 __________________ keywords: computational thinking, requisite skills, research, knowledge science, vuca. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(1) (2023) 11-22 https://doi.org/10.34010/injiiscom.v4i1.9558 mailto:s2220401@jaist.ac.jp adam mukharil bachtiar. computational thinking: the essential skill …| 12 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 khare, 2016). the extra challenge was added to society as the effect of this concept. the vuca world offered the uncertainty and unpredictable future that break in the stability of the digital era (johansen & euchner, 2013). the vuca presence threatens the process of creating innovation in the digital platform. society 5.0 has stated the requisite skills for humans in the 21st century as the primary skills to be achieved by undergraduate or graduate students. communication, critical thinking, collaboration, creativity, and problemsolving skills are some powerful words that appear in the list of requisite skills for the 21st century (gutiérrez-núñez et al., 2022; semsri et al., 2022; van laar et al., 2020). but the uncertainty word needs to be satisfied and force society to add extra skill as the complement. in the middle of the chaos, knowledge science became more mature nowadays as one of the knowledge domains. the hard and soft research type was published to encourage the role of knowledge science in the higher education domain (huang et al, 2016). without giving enough space for debating about the position of knowledge science compared with the existing knowledge domain, many researchers in the field have successfully filled the gap in the research in information science. the granularity between information and knowledge creates a summary that knowledge science is well deserved as the new domain (zins, c. 2006). like the other field, knowledge science needs to answer the challenge given by the vuca world. future researchers in this field should own the extra skills, excluding the essential skills in the 21st century, to survive and continue to produce the following research. some recent research shows the appearance of a new skill called computational thinking. most studies reveal the importance of computational thinking in education, including the knowledge science domain. that thinking method successfully demonstrated the change in human behaviors in solving the problem (sermsri et al, 2022; kong, s. c. 2022; yasar et al, 2023). the way of teaching is also got affected by computational thinking. today’s scientists not only leverage computational tools to conduct their investigations but often must also contribute to designing the computational tools for their specific research (hurt et al., 2023). because of this need, this study is focused on revealing the intuition and the way of thinking of computational thinking in conducting knowledge science research. using descriptive analysis, this study intuitively examined each element of computational thinking in some knowledge science research to be a general overview for future researchers in the knowledge science field. 2. method this study involves two significant concepts: computational thinking and knowledge science. both seem dif ferent, but computational thinking is a general thinking skill that can exist in every domain. the first sub-section will focus on explaining computational thinking, followed by the second, which focuses on explaining knowledge science as the domain. https://doi.org/10.34010/injiiscom.v4i1.9558 13 | international journal of informatics information system and computer engineering 4(1) (2023) 11-22 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 2.1. computational thinking jeannette wings, the founder of computational thinking (also known as ct), used words such as problem-solving method and how computers execute the solution to give a powerful understanding of computational thinking (wing. j. m., 2014). there is one famous statement from her that if one person thinks using computational thinking, then the person will involve in formulating problems and their solutions so that the solutions are represented in a form that an information-processing agent can effectively carry out (shute & asbellclarke, 2017). the word “an information-processing agent” refers to the computer. some research demonstrates how computational thinking is the essential thinking method across many fields. tables 1 and 2 show the research on computational thinking in some science fields, either natural or social science. table 1. computational thinking in some science research study explanation qin, h., 2019 this study comprehensively develops how ct helps the participants learn about bioinformatics using computer laboratory exercises. the researchers can examine how to implement ct early in bioinformatics learning even though they cannot determine which elements are significant weintrop et al., 2016 this study argued about the position of ct in supporting the mathematics and science context. the modeling and simulation became the most significant part affected by ct chongo et al., 202 chemistry also became one field that ct invaded. the experiment using plugged and unplugged ct method is the central part of this study güven & gulbahar, 2020 social science will be the last field to be predicted as ct invades. this study provided an excellent comprehensive about how to implement ct in social studies table 2. four elements of computational thinking (mack & khare, 2016) element of ct explanation abstraction a. determine the fundamental problem from all the phenomena. b. reformulate into solvable and can be familiarized as the computational case decomposition break down the problem into several sub-problems that can be more solvable intuitively algorithm construct a series of the structured process to be followed in solving the problem pattern recognition finding the similarity and shared characteristics between the problem https://doi.org/10.34010/injiiscom.v4i1.9558 adam mukharil bachtiar. computational thinking: the essential skill …| 14 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 those studies well-explained that ct can be a mature thinking skill in a short time after being declared by the founder. to think using this approach, the researchers should implement four elements of ct. with those elements, ct is possible to be implemented. 2.2. knowledge science based on nakamori (2011), knowledge science is an emerging discipline resulting from the demands of a knowledge-based economy and information revolution. the diversity between information and knowledge triggered the shift of the field's name from information science to knowledge science. changing the area's name reflects that current information science primarily focuses on exploring the mediating aspects of human knowledge (zins, c, 2006). unlike information science, which focuses on manipulating the form and structure of information, knowledge science concentrates on optimizing the knowledge creation process either by producing new knowledge using some methodologies or serving the optimization of human and social concepts in the knowledge creation process. there are two classifications of research in knowledge science (huang et al., 2016; hlupic et al, 2002). the classification is based on the type of process in the knowledge management area, which is a significant area in the knowledge science field. figure 1 shows the research classifications in the knowledge science area. both types share the central role of knowledge science, such as knowledge creation, knowledge sharing, knowledge management, and knowledge evaluation equall. fig. 1. two classifications in knowledge science research https://doi.org/10.34010/injiiscom.v4i1.9558 15 | international journal of informatics information system and computer engineering 4(1) (2023) 11-22 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 3. results and discussion 3.1. examples of hard and soft-type research some examples of each type of knowledge science research were represented in this study before mapping each element of ct into the research. based on the examples, the intuition for making differences between hard and soft-type research in knowledge science can be understood. table 3 shows some examples of hard-type knowledge science research in the school of knowledge science, jaist. table 4 shows the opposite of the hard-type research, which is soft-type research in the knowledge science domain. table 3. examples of hard-type knowledge science research research short summary ono et al, 2022 this research produces technology for skiers learning using virtual reality. deep learning was used to recognize the skiing posture to be evaluated tan et al, 2019 this research produces knowledge in the form of infographics about catalyst degradation mechanisms based on operand spectro imaging and unsupervised learning from 3d images. hamanaka et al, 2016 this research focused on implementing lerdahl and jackendoff’s (1983) generative theory of tonal music (gttm) to generate new music based on the training data miyata et al., 2012 this research generates several procedural technologies that can be used to generate pattern images (3d models). torii et al., 2022 this research predicts movement characterizes the degree of animacy and measures it using granger causality. table 4. examples of soft-type knowledge science research research short summary sinthupundaja et al, 2019 this research examined the importance of the causal combinations of knowledge-acquisition conditions using fuzzy set qualitative comparative analysis. shahzad et al, 2016 this research aimed to identify if integration between knowledge strategy and knowledge management (km) processes leads to organizational creativity and performance. hashimoto, 2006 this research focused on modeling to clarify the evolutionary process of language, and evolutionary economics defines the dynamics of economic phenomena. uchihira et al, 2012 this research generated a model to optimize the knowledge transfer process in r&d project management. kim, 2017 this research is aimed to identify the factors that influence the creation of innovative ideas. the two workshops were conducted to reveal influential factors. https://doi.org/10.34010/injiiscom.v4i1.9558 adam mukharil bachtiar. computational thinking: the essential skill …| 16 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 3.2. abstraction for simplification the intuition behind abstraction is to determine the essential part of the problem and generalize the problem to find its proper form. it can simplify the complex problem to become an intuitive and identified problem. some unnecessary elements of the problem can be excluded so that the focus of researchers increases simultaneously. for example, there are studies by hamanaka et al., 2016 and ohmura et al. about generating music based on the relationship using lattice probability distribution (ohmura et al, 2018). that study found the essence of the relationship between the tonal in one music. then, the relationship was used to generate new music. figure 2 is the overview of the abstraction in the research. in the soft-type, abstraction can be used for generalizing the procedure and its problem into one model. then, this model will be improved during the research. the famous abstraction result in soft-type research is the seci model that focuses on optimizing the knowledge creation process in one organization (farnese et al, 2019). 3.3. decomposition for reducing the complexity the intuition behind the decomposition is breaking down the identified problem into several easy-to-chunk problems. the primary strategy is about divide and conquers, which will lead the researchers to the estimated solution. in some research, decomposition is not easy, especially when the problem involves one complex system. soft system methodology and i-system can decompose the complexity among the constituent party followed by their emergence (nakamori, y, 2011; mingers et al, 1992). the decomposition in knowledge science research can be used to examine the interventions in one isystem for later, the solution will be recognized by the three dimensions, such as scientific, collaboration, and creative dimension. in the end, the three solutions offered by each dimension will be integrated in the final phase of the isystem. fig. 2. the abstraction of tonal music generation https://doi.org/10.34010/injiiscom.v4i1.9558 17 | international journal of informatics information system and computer engineering 4(1) (2023) 11-22 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 the study by kim, e., 2017 has successfully demonstrated how the decomposition worked well. the study is focused on revealing the influential factors in idea generation and enhancing them using analogical thinking. the experiment in that study was divided into two workshops. the first workshop focused on revealing the influential factors, and, in the end, three influential factors were revealed. then, the second workshop focused on enhancing those factors using analogical thinking. the conclusion in one research can be achieved through several processes. each process produces the output represented as the input in the following procedure. the properness to break down the problem, especially to be some processes that can be parallelized, will increase the efficiency and optimal level to achieve successful research. 3.4. algorithm to lead the research the algorithm is a very familiar element among all the ct’s elements. it is mandatory for the researchers to build a structured and sequenced process to lead the problem into the solution. in knowledge science research, the algorithm can play a role as the methodology in one research. the study from miyata, k., 2012 demonstrated the algorithm in the form of procedural technology for pattern generation or 3d pattern generation. the pattern can be used in a kimono or building structure. step-by-step how the procedural technology was constructed from the actual pattern is one clear example of how the algorithm took an essential part in this research. another example algorithm can play a role as a procedure about how to conduct the experiment workshop. uchihira et al., 2012 experimented with making a model for optimizing the knowledge transfer process in research and development projects. the algorithm helped the study to illustrate the flow in a structured project case and to conduct an internalization workshop that consisted of six steps. each step is well-structured. the algorithm is about not only structuring the programming process but also the experimental process. 3.5. pattern recognition for finding the similarity this element plays a significant role in satisfying the uncertainty. rather than finding similarities, some researchers often focus on finding the differences among the research. most of the study is excellent in generating new approaches or results. even though they seem to be different, they have connected to each other. the five examples of hard-type knowledge science research in table 3 focus on finding the hidden knowledge using several methods in knowledge discovery methodology. the differences are in the source and form of the knowledge. similar connecting lines also happen in soft-type knowledge science research. the experimental method and the proposed model are the shared characteristic between studies. the differences in the domain and the design of experiments. using the capability for finding the similarity can help the researchers to shorten the time for getting the intuition behind the research. from the similarity, they can mark the area in the research domain map that has already been invaded by the other researchers and find the gap between them. table 5 shows similarities in some research in tables 3 and 4. https://doi.org/10.34010/injiiscom.v4i1.9558 adam mukharil bachtiar. computational thinking: the essential skill …| 18 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 table 5. similarities between the research on the school of knowledge science in jaist with other research outside jaist research result of research similar research as an evidence ono et al, 2022 the proper interaction model for vr to help novice skiers similar research was found in creating an interaction model for elder skiers using vr technology. the recent study involved deep learning in evaluating the properness of body posture when skiing. tan et al, 2019 the similarity concepts and infographics that were produced by the unsupervised learning in the experimental material several concepts of data analytics and machine learning can be applied in material science. unsupervised learning was found inone study to be a knowledge discovery process in grouping microstructure material. the experimental object and the intention of data analytics can be a differentiating factor between the studies. hamanaka et al, 2016 the relationship rules between the tune in music for generating new music. the vertical rule is an outstanding result of this research. there is similar research about the vertical pattern in music and how to discover it. the discovery process is called computational music analysis. even though the similarity was so high, the difference is in the form of explicit knowledge produced by the algorithm. torii et al, 2022 the pattern of movement characteristics was measured by the degree of animacy and granger causality for the robotic domain the collision prediction from the robotic movement scenario also resulted from another research. the subdomain from the studies is different, and the focus of movement prediction can differ from the studies. sinthupund aja et al, 2019 the concept of the causal combinations of knowledgeacquisition condition rather than using fuzzy logic, the other study used bayesian network as their primary method to reveal the causal combination of the knowledge-acquisition condition. the dissimilarity also can be found in the proposed concept of knowledge acquisition. shahzad et al, 2016 the validated research model of the hypothesis about the integration between knowledge strategy and nowledge management and its correlation to organizational creativity and performance there are some studies about integrating other possible factors into knowledge management strategy. this further study focused on integrating the aspect of intellectual capital into knowledge management. the dissimilarity factors are the proposed integrated factors, and the destination of the effect comes from the integration procedure. hashimoto, 2006 the new pattern when doing recursion is to make the hierarchical structure one research mentioned several patterns in the linguistic domain. both studies are about finding a pattern in the linguistic model, but the methods used are different and also for their intention. 4. conclusion from the revealing process of ct in some knowledge science research, there are some conclusions for this research, such as: (a) computational thinking is a complementary skill to 21st-century skills. (b) the primary elements of ct in knowledge science research are abstraction and pattern recognition. the other two elements are similar to other elements in different skills. (c) abstraction optimizes how knowledge science researchers generalize problems and makes a model from this. (d) pattern recognition focuses on finding the https://doi.org/10.34010/injiiscom.v4i1.9558 19 | international journal of informatics information system and computer engineering 4(1) (2023) 11-22 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 similarity among the studies so that the researchers can focus more on dissimilarity factors. references ashrafi, m., davoudpour, h., & khodakarami, v. (2017). a bayesian network to ease knowledge acquisition of causal dependence in cream: application of recursive noisy‐or gates. quality and reliability engineering international, 33(3), 479-491. carayannis, e. g., & morawska-jancelewicz, j. (2022). the futures of europe: society 5.0 and industry 5.0 as driving forces of future universities. journal of the knowledge economy, 1-27. chongo, s., osman, k., & nayan, n. a. (2021). impact of the plugged-in and unplugged chemistry computational thinking modules on achievement in chemistry. eurasia journal of mathematics, science and technology education, 17(4). conklin, d. (2002, august). representation and discovery of vertical patterns in music. in music and artificial intelligence: second international conference, icmai 2002 edinburgh, scotland, uk, september 12–14, 2002 proceedings (pp. 32-42). berlin, heidelberg: springer berlin heidelberg. de vries, a. w., faber, g., jonkers, i., van dieen, j. h., & v erschueren, s. m. (2018). virtual reality balance training for elderly: similar skiing games elicit different challenges in balance training. gait & posture, 59, 111-116. farnese, m. l., barbieri, b., chirumbolo, a., & patriotta, g. (2019). managing knowledge in organizations: a nonaka’s seci model operationalization. frontiers in psychology, 10, 2730. gutiérrez-núñez, s. e., cordero-hidalgo, a., & tarango, j. (2022). implications of computational thinking knowledge transfer for developing educational interventions. contemporary educational technology, 14(3), ep367. güven, i., & gulbahar, y. (2020). integrating computational thinking into social studies. the social studies, 111(5), 234-248. hamanaka, m., hirata, k., & tojo, s. (2016). implementing methods for analysing music based on lerdahl and jackendoff’s generative theory of tonal music. computational music analysis, 221-249. hashimoto, t. (2006). evolutionary linguistics and evolutionary economics. evolutionary and institutional economics review, 3, 27-46. hlupic, v., pouloudi, a., & rzevski, g. (2002). towards an integrated approach to knowledge management: ‘hard’,‘soft’and ‘abstract’issues. knowledge and process management, 9(2), 90-102. huang, f., gardner, s., & moayer, s. (2016). towards a framework for strategic knowledge management practice: integrating soft and hard systems for competitive advantage. vine journal of information and knowledge management systems. https://doi.org/10.34010/injiiscom.v4i1.9558 adam mukharil bachtiar. computational thinking: the essential skill …| 20 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 hurt, t., greenwald, e., allan, s., cannady, m. a., krakowski, a., brodsky, l., ... & dorph, r. (2023). the computational thinking for science (ct -s) framework: operationalizing ct-s for k–12 science education researchers and educators. international journal of stem education, 10(1), 1-16. johansen, b., & euchner, j. (2013). navigating the vuca world. research-technology management, 56(1), 10-15. kemmerer, d. (2012). the cross‐linguistic prevalence of sov and svo word orders reflects the sequential and hierarchical representation of action in broca ’s area. language and linguistics compass, 6(1), 50-66. kitahara, a. r., & holm, e. a. (2018). microstructure cluster analysis with transfer learning and unsupervised learning. integrating materials and manufacturing innovation, 7, 148-156. kong, s. c. (2022). problem formulation in computational thinking development for nurturing creative problem solvers in primary school. education and information technologies, 1-20. li, c., & bacete, g. (2017). international journal of innovation studies. in workshop design for enhancing the appropriateness of idea generation using analogical thinking (vol. 1, no. 2, pp. 134-143). mack, o., & khare, a. (2016). perspectives on a vuca world. managing in a vuca world, 3-19. mingers, j., & taylor, s. (1992). the use of soft systems methodology in practice. journal of the operational research society, 43(4), 321-332. miyata, k. (2012, november). procedural technology and creativity mining. in 2012 seventh international conference on knowledge, information and creativity support systems (pp. 169-174). ieee. nakamori, y. (ed.). (2011). knowledge science: modeling the knowledge creation process. crc press. ohmura, h., shibayama, t., hirata, k., & tojo, s. (2018). music generation system based on human instinctive creativity. proceedings of computer simulation of musical creativity (csmc2018). ono, s., kanai, h., atsumi, r., koike, h., & nishimoto, k. (2022, june). learning support and evaluation of weight-shifting skills for novice skiers using virtual reality. in adaptive instructional systems: 4th international conference, ais 2022, held as part of the 24th hci international conference, hcii 2022, virtual event, june 26–july 1, 2022, proceedings (pp. 226-237). cham: springer international publishing. qin, h. (2009, march). teaching computational thinking through bioinformatics to biology students. in proceedings of the 40th acm technical symposium on computer science education (pp. 188-191). sermsri, n., sukkamart, a., & kantathanawat, t. (2022). thai computer studies student teacher complex problem-solving skills development: a cooperative https://doi.org/10.34010/injiiscom.v4i1.9558 21 | international journal of informatics information system and computer engineering 4(1) (2023) 11-22 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 learning management model. journal of higher education theory and practice, 22(16), 87-99. shahzad, k., bajwa, s. u., siddiqi, a. f. i., ahmid, f., & raza sultani, a. (2016). integrating knowledge management (km) strategies and processes to enhance organizational creativity and performance: an e mpirical investigation. journal of modelling in management, 11(1), 154-179. shin, j. c. (2015). mass higher education and its challenges for rapidly growing east asian higher education. mass higher education development in east asia: strategy, quality, and challenges, 1-23. shute, v. j., sun, c., & asbell-clarke, j. (2017). demystifying computational thinking. educational research review, 22, 142-158. sinthupundaja, j., chiadamrong, n., & kohda, y. (2019). internal capabilities, external cooperation and proactive csr on financial performance. the service industries journal, 39(15-16), 1099-1122. tan, y., matsui, h., ishiguro, n., uruga, t., nguyen, d. n., sekizawa, o., ... & tada, m. (2019). pt–co/c cathode catalyst degradation in a polymer electrolyte fuel cell investigated by an infographic approach combining three dimensional spectroimaging and unsupervised learning. the journal of physical chemistry c, 123(31), 18844-18853. torii, t., oguma, k., & hidaka, s. (2022). animacy perception of a pai r of movements under quantitative control of its temporal contingency: a preliminary study. artificial life and robotics, 27(3), 448-454. uchihira, n., hirabayashi, y., sugihara, t., hiraishi, k., & ikawa, y. (2012, july). knowledge transfer in r&d project management: application to businessacademia collaboration project. in 2012 proceedings of picmet'12: technology management for emerging technologies (pp. 3473-3480). ieee. van laar, e., van deursen, a. j., van dijk, j. a., & de haan, j. (2020). de terminants of 21st-century skills and 21st-century digital skills for workers: a systematic literature review. sage open, 10(1), 2158244019900176. wang, y., ye, x., yang, y., & zhang, w. (2017, november). collision-free trajectory planning in human-robot interaction through hand movement prediction from vision. in 2017 ieee-ras 17th international conference on humanoid robotics (humanoids) (pp. 305-310). ieee. weintrop, d., beheshti, e., horn, m., orton, k., jona, k., trouille, l., & wilensky, u. (2016). defining computational thinking for mathematics and science classrooms. journal of science education and technology, 25, 127-147. wiig, k.m.: integrating intellectual capital and knowledge management. long range planning. 30, 399–405 (1997). https://doi.org/10.1016/s00246301(97)90256-9. wing, j. m. (2014). computational thinking benefits society. 40th anniversary blog of social issues in computing, 2014, 26. yasar, o., maliekal, j., veronesi, p., little, l., meise, m., & yeter, i. h. (2023). retrieval practices enhance computational and scientific thinking skills. stem, steam, https://doi.org/10.34010/injiiscom.v4i1.9558 adam mukharil bachtiar. computational thinking: the essential skill …| 22 doi: https://doi.org/10.34010/injiiscom.v4i1.9558 p-issn 2810-0670 e-issn 2775-5584 computational thinking and coding: evidence-based research and practice in children’s development, 16648714, 142. zins, c. (2006). redefining information science: from “information science” to “knowledge science”. journal of documentation. 62, 447–461 (2006). https://doi.org/10.34010/injiiscom.v4i1.9558 93 | international journal of informatics information system and computer engineering 3(1) (2022) 93-104 design and implementation of a cloud based decentralized cryptocurrency transaction platform benjamin kommey*, eric tutu tchao, emmanuel osae-addo, asiedu biney kofi yeboah, derick biney kwame nkrumah university of science and technology, kumasi, ghana *corresponding email: bkommey.coe@knust.edu.gh a b s t r a c t s a r t i c l e i n f o trading in the crypto-currency market has seen rapid growth and adoption, as well as the interest in crypto related technologies like blockchain and smart contracts. smart contracts have gained popularity in building so called decentralized applications (dapps) and decentralized finance (defi) apps, mainly because they are more secure, trustworthy, and largely distributed (removes centralized control). defi applications run on the blockchain technology and are secured by blocks (nodes) connected by cryptographical hash links. defi applications have a great potential in the crypto-currency trading domain, providing more secure and reliable means of trading, and performing transactions with crypto-currencies. only verified transactions are added to the blockchain after being approved by miners through a consensus mechanism and then it is replicated (distributed) among the nodes on the blockchain network. this research paper proposes a defi crypto exchange by integrating a numerous-signature stamp with a crypto api. a numerous-signature stamp solves the issue of transaction verifiability and authenticity. a crypto api provides the data about each crypto currency with which trades and transactions will be performed. this paper also discusses the technical background of the technology and a few related works. decentralization of transactions through smart contracts on the blockchain will improve trust, security and reliability of transactions and trades. article history: received 25 may 2022 revised 30 may 2022 accepted 10 june 2022 available online 26 june 2022 aug 2018 __________________ keywords: trading, cryptocurrency, defi, finance, technology, applications. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 93-104 mailto:bkommey.coe@knust.edu.gh benjamin kommey et al. design and implementation of a cloud based | 94 1. introduction a defining feature of cryptocurrencies is that they are generally not issued by any central authority, rendering them theoretically immune to government interference or manipulation. the ghana cedis is a one of many currencies that are backed by the us dollars however due to imminent economic decline, the cedi value has depreciated considerably prior to its inception. the value of the ghana cedis (like any other fiat currencies) has a positive correlation with the trust of the people in their government and financial system so deductively if the people’s trust dips dramatically; it could lead to hyperinflation of the cedi. to cite a common example, the zimbabwean dollar due to hyperinflation made their currencies more worthless than the paper they were printed on (toby, 2021). states that the african continent being the prominent user of mobile cash transactions has made advancement in modernization of finance system. mobile money transaction boomed globally in 2020, especially in sub-saharan africa which accounted for 43% of all new accounts, according to the gsm association. more than half of such accounts are in africa, which has been the fastest-growing region for mobile phone growth for several years. however, the security of such a system is flawed in a few ways making it the scavenging grounds of fraudsters. the use of just a mobile phone and sim card being the innovation for its convenience is also the perpetuator of exploitation. if the sim cards can be easily identified and targeted, there always be loopholes to scam people of their hard-earned money. in ghana, there are at least three major telecommunications companies all having their own mobile money system. ensuring to a degree that the system is not monopolized by one, but this leads to another inconvenience. in a region where there are agents of only one mobile money system say mtn, the users of other networks would not have access to their money. furthermore, the transfers of cash between different networks, though not a herculean task for many, is far from seamless and incur extra charges. the human complexity of cryptocurrency apps has made cryptocurrency the playground of a select few. even in far more developed countries the average person must be guided in cryptotrading because most applications have so many technicalities in their usage. with ghana’s high illiteracy rate (and even high financial illiteracy rate), we are not an exception (mukhopadhyay, 2016). the ghana cedis and mobile money systems are limited in the sense that neither one can ensure a secure and fast money transfer to anyone around the world at any time. the use of cryptocurrency system provides protection against inflation caused by government’s poor management of fiat currencies. a fast majority of the world are slowly embracing the system making it worth a market cap of trillions which is a testament to their faith in them. this peerto-peer system cuts off the necessity for the bank and government’s involvement. most cryptocurrencies are mined using brute force algorithm so that the number of blocks mined per day remains approximately constant to control the rate of introduction of new currencies preventing any excess influx of the currencies that can lead to inflation. 95 | international journal of informatics information system and computer engineering 3(1) (2022) 93-104 moreover, cryptocurrencies are secured by cryptography, which makes it nearly impossible to counterfeit or doublespend. every transaction must be verified by thousands of computers around the world. it gives the user pseudonymity which makes it more secure than the mobile money system. they are utilized for cross-border transactions to a limited extent. an excellent illustration of such decentralized transfers is flash loans in decentralized finance. these loans can be executed instantly and are used in trading because they are done without supporting collateral. cryptocurrency is the new wave that the world is riding on, an innovative technology which has the prospect to dethrone the current finance system and if care not taken ghana will be left behind. this project was proposed to facilitate secure crypto transactions and trading. trading involves the buying and selling activity between a buyer and a seller. the commodity that the two parties can exchange are such as fiat, stocks, bonds, or crypto currency. crypto currency is a digital currency or digital asset that is decentralized and not regulated by any bank. the decentralization of cryptocurrencies is based on blockchain technology, which is a distributed ledger system where each node is linked together in a peer-to-peer (p2p) network. each node in the network owns a similar copy of the ledger or transaction, which is verified and synchronized with the creation of a new block through a consensus protocol. the consensus protocol eliminates the need for a centralized or trustworthy entity, such as a bank or government agency (uriawan, 2020). after reaching an alltime high of $3.3 trillion on december 29, 2021, the global crypto market capitalisation was $2.21 trillion (coinmarketcap). the value of the cryptocurrency market has increased by a factor of six since november 2020, when it was only slightly more than $578 billion. in may 2021, the average daily trade volume reached a peak of $500 billion before levelling out at $120 billion (statista). due to its long history of increasing in value, investors continue to have faith in bitcoin, the first decentralized digital currency underpinned by blockchain technology. with a market cap of nearly $1 trillion, bitcoin holds a market dominance of 40% as of december 2021, followed by ether, which benefits from a 20% dominance, and other altcoins (coins other than bitcoin) such as solana and xrp. the number of blockchain wallet (instrument needed by crypto owners to store and manage their crypto) users went from 0 to 80 million in the past 10 years (statista) this gives a clear snapshot of the increase in popularity of cryptocurrency as a form of payment (puspitawati, 2021). the underpinning blockchain technology of cryptocurrencies inherently allows for fast, secure, and tamper-proof recording of transactions. nonetheless, cryptocurrencies aren't immune to cybercrime, crypto-thefts, and frauds. in 2021, global crypto thefts accounted for a loss of $681 million. the weak link in this system is largely due to the exchange or trading platform, which mostly are built on a web 2.0 architecture using centralized protocol. this architecture makes lots of exchange platforms still vulnerable by design to account for hacking and false transaction scams, quite like any other web application (rahayu, 2022). the purpose of this paper is to discuss the context of smart contracts and benjamin kommey et al. design and implementation of a cloud based | 96 how they can be implemented to develop a safer and secure decentralized platform for trading and exchanging decentralized digital assets like cryptocurrencies to smoothly usher us into the age of cashless system and provide a more secured means of exchanging, accessing, and trading cryptocurrencies. 2. related works this section gives a chronological and thorough condensation of cryptocurrencies from the time of its conception to the current trends and evolutionary milestones. dating back to 1980s where the first attempts started, then to the first token currency of 90s and eventually the blockchain technology together with its derivatives. one of the first attempts at cryptocurrencies occurred in the netherlands in the late 1980s. a batch of developers sort to link money to smart cards designed to cater for night-time thefts on petrol stations. vehicle drivers would use these cards as a means of transaction instead of cash, leaving no paper monies around for thieves. around that time, david chaum, an american cryptographer experimented on another form of electronic cash. he imagined a token currency which could be sent among people privately and securely. he developed a formula he called the “blinding formula” which would be used to encrypt information transferred between individuals. the “blinded cash” as he would call it could be transferred among individuals who would be having a signature of authenticity. he went on to establish digicash where he would put his idea into practice. though his company went bankrupt, his concepts and formulas of encryption played key roles in the development of subsequent currencies. start-ups began making efforts to further the goals of digicash in the 1990s. companies like paypal which is likely the company with the largest lasting impact on the financial world were created. individuals could send and receive money over a web browser quickly and securely. it inspired other startups like egold which attempted to create a platform where precious metals like gold could be traded. e-gold gave individuals online credit in exchange for physical gold and other precious metals. it was shut down eventually due to scams and other issues (digital curency, 2007). in 1998, wei dai, a computer engineer and graduate of the university of washington first revealed b-money. it was purposed to be a distributed electronic cash system which would be anonymous. wei described b-money as “a scheme for a group of untraceable digital pseudonyms to pay each other with money and to enforce contracts amongst themselves without outside help”. although bmoney was never officially launched, bmoney endeavoured to render many of the services offered by cryptocurrencies today (buntinx, 2016). around the same period, nick szabo created bit gold. bit gold came with a proof-of-work system that mirrors bitcoin mining process in certain ways. proof-of-work is a consensus mechanism that is used to confirm and record cryptocurrency transactions. it’s a means of adding new blocks of transactions to a cryptocurrency blockchain. it involves generating hash codes that would have to match the target hash code for the block. szabo’s bit gold had its most revolutionary aspect to be its decentralized status. thus, bit gold sort to avoid reliance on centralized and 97 | international journal of informatics information system and computer engineering 3(1) (2022) 93-104 authorities. ultimately, bit gold also proved unsuccessful as b-money but gave inspiration for future digital currencies. as one of the most successful prebitcoin digital currencies, hashcash was also developed i the mid-1990s. it was developed for purposes of minimizing email spam and preventing ddos attacks. hashcash also used a proof-of-work algorithm which would aid the generation and distribution of new coins like modern cryptocurrencies. just like previous developments, hashcash also became less effective due to increased need for processing power though most of its elements were used in the development of bitcoin (griffith, 2014). a blockchain is naturally a network of connected computer systems that duplicates and distributes a digital ledger of transactions. a blockchain divides its data into temporally and cryptographically linked blocks. a blockchain is often a sort of database that only allows for reading and adding. the peer-to-peer network nature of blockchain is a result of its decentralized architecture. as a result, users (peers) communicate with one another directly without the aid of authorities or other trustworthy intermediaries. the blockchain technology was used to implement bitcoin and other contemporary cryptos. many blockchain applications have been developed over the years and have revolutionised the way people view digital currencies. commonly cited applications include using digital assets on a blockchain to represent custom currencies and financial assets, the ownership of an underlying physical device, non-fungible assets such as domain names and more advanced applications such as decentralized exchange among others. another important area of the blockchain technology is the use of smart contracts. these are systems which automatically move digital assets according to arbitrary pre-specified rules. for example, one might have a treasury contract of the form "a can withdraw up to x currency units per day, b can withdraw up to y per day, a and b together can withdraw anything, and a can shut off b's ability to withdraw". when satoshi first established bitcoin in january 2009, he was simultaneously coining two radical and untested concepts (nakamoto, 2008). the first is the bitcoin, a decentralized online currency which is peer-to-peer and maintains a value without any backing, intrinsic value, or central issuer. so far, bitcoin as a currency has taken up majority of the public attention, in terms of both the political aspects of a currency without a central bank and its extreme uncertain volatility in price. however, there is also a different, equally important, aspect of satoshi's grand experiment. thus, the concept of a proof of work-based blockchain that allows for public consent on the issue of transactions. bitcoin can be described as a first-to-f system. thus, if an individual has 60 btc, and simultaneously sends the same 60 btc to entity a and to entity b, only the transaction that gets validated first will be processed. many cryptocurrencies of modern days then started emerging using the concept of bitcoin. in a paper by jaehyung lee and minhyung cho in 2018, exeum was introduced, a decentralized architecture that issued pegged token backed by world assets, including fiat currencies. pegged tokens are over-collateralized by the virtual assets exchanged in the decentralized virtual asset trading provided by the structure, effectively remoulding the price stability dilemma into maintaining the disparity between benjamin kommey et al. design and implementation of a cloud based | 98 the virtual asset exchange and real-world exchanges. the system implemented several mechanisms to motivate market makers and preserve the peg – a rebate for maker orders, the swap rate adjusted based on the demand of the asset in the exchange, and loose protection of the peg by the market maker dapp and the initial reserve. the exeum project to democratize market making activity by enroling arbitrage miners, using the market making software provided by exeum (lee, 2018). in a 2018 paper, chi ho lam disclosed a system to support a p2p cross chain crypto asset exchange based on signature scheme to facilitate a p2p crosschain crypto asset exchange. the system provides a universal secure and direct way for traders to exchange crypto assets across different chains without hassle. the benefit of this mechanism was that it applied to the signature level instead of the protocol level (lam, 2019). a study in 2019 by stanislaw dro˙zd˙z, and his colleagues from complex systems theory department, institute of nuclear physics, polish academy of sciences, provided unwavering support for the hypothesis of the gradual development of a novel and partially independent market, synonymous to the worldwide foreign exchange (forex) market, wherein cryptocurrencies are traded in a free-standing manner. in more practical terms, this meant that not only the bitcoin but even the whole emerging crypto market may, eventually, offer ’a hedge or safe haven’ for currencies, gold, and commodities. in a 2020 paper mohd faizal yusof and his colleagues aimed to clarify how to implement a cryptocurrencies payment platform which comprised of a web component to allow end-users to declare cryptocurrencies owned by them, a mobile component to support end-users who prefer mobile phones and a backend system to manage the collection of zakat in cryptocurrencies and integration with the entire system (yusof, 2021). stefan ciberaj and martin toma´sek in their ‘crypto trading platform's article included a prototype implementation of a proof-of-concept architecture for a bitcoin trading platform. an android client with a focus on user experience (ux) and an application programming interface (api) the client utilizes that is also directly accessible to users make up the offered interfaces. the server leverages cloud computing design patterns and is made up of microservices. it has a trio-tier architecture with a focus on scalability (fang, 2022). another 2021 study by method and system for crypto-asset transfers were introduced by berengoltz and his associates. the method includes sharding a wallet private key so that each shard is given to a different secure module, generating signatures by each secure module based on a respective shard of the sharded wallet private key and obtained trading platform credentials, and verifying the cryptoasset transaction when a threshold of the generated signatures is found to match (brenglotz, 2021). an issue with a sharding-based strategy is the security worry that develops when a shard is hacked, leading to shard takeovers where one shard attacks another and information is lost (frankenfield, 2021; presthus, 2017; binns, 2022) (fig. 1). 3. method 3.1. system architecture 99 | international journal of informatics information system and computer engineering 3(1) (2022) 93-104 fig. 1. system architecture the system architecture is as depicted in fig. 1. a user accesses or enters the url of the decentralized trading platform (e.g., slimetrader.com) which displays the user interface (ui). the static files for the ui are retrieved via ipfs (a decentralized off-chain storage solution). when the user initiates a transaction, forms are provided for the details of the transaction to be entered. transaction details are commodity / cryptocurrency to transfer, recipient address and amount. on clicking submit, the transaction is encrypted or signed with the user’s private key by meta mask (all write actions must be signed, otherwise the transaction will be rejected by the nodes on the blockchain). providers like meta mask offer nodes that allows the user / frontend to connect and interact with the blockchain. the smart contract consists of the business logic that automatically processes the transaction details in the ethereum virtual machine and checks if sender balance is enough for the transaction or debit the sender’s account or verify the recipient’s wallet address or credit the recipient’s address with the debited amount. the nodes verify the transaction via a consensus protocol, once approved the transaction is then hashed as a block onto the blockchain (as a result no 3rd party or central authority is required to provide trust). this data stored in the blockchain is queried and sent to the frontend to be displayed. 3.2. system block diagram the cryptocurrency platform, in this case represented with a block diagram, as illustrated in fig. 2 contains mainly four (4) blocks namely the user block, frontend block, provider block and ethereum blockchain block. the procedures for transaction are as explained in subsection titled “system architecture”. fig. 2. system block diagram 3.3. system workflow the system workflow is as shown in fig. 3. user loads application in an internet browser and logs in / signs up using a meta mask wallet account. if a user does not have a meta mask account, direct the user to the meta mask page to acquire one. on logging in, the dashboard is loaded to show transactions and your account holdings. if the user wants to make a transaction, the user clicks on the send button and a form is displayed and the user can fill in details of the transaction which includes the name of receiver and amount to be transferred. after which, the user clicks on the submit/make transfer button to initiate the transaction. a confirmation popup message is displayed for the user to verify the transaction and if the user confirms, the transaction is made, and the user is redirected to the dashboard. current holdings are displayed on the dashboard. but if the user’s wallet is credited, a notification appears as a popup. the user benjamin kommey et al. design and implementation of a cloud based | 100 clicking on the notification will be directed to the dashboard to view the update. fig. 3. system workflow 3.4. system software design the system software was designed using modular method. the modules or software components are frontend, login, transitions and backend and this is illustrated graphically in fig. 4 and detailed descriptions are given in table 1. fig. 4. system software design diagram table 1. software design components and description frontend development next.js next js is a framework built on react js which is a javascript framework designed to be declarative, component-based, and portable. react makes it easier to create interactive user interfaces. react efficiently updates and renders just the right components when data changes after designing simple views for each state in an application. since the logic of a component is written in javascript instead of templates, you can easily pass data through your app and keep state out of the dom. sass sass is the most mature, stable, and powerful professional css extension language in the world. sass boasts of more features and capabilities than any css extension language out there and that is why we opted for it in our project. swipe.js swiper is a modern mobile touch slider with hardware accelerated transitions and amazing native behaviour. it is purposed to be used in building mobile websites, mobile web apps, and mobile native or hybrid applications. swiper, along with other great components, is a part of framework7, a fully featured framework for 101 | international journal of informatics information system and computer engineering 3(1) (2022) 93-104 building ios & android app. this helps to comply with a multi-platform compatibility like digi crypto. swiper comes with a very resourceful api. it allows for the creation custom paginations, navigation buttons, parallax effects and other vast options. log in meta mask to help secure and useable ethereum-based websites, meta mask was developed. it specifically takes care of account administration and establishing a user's connection to the blockchain. users who already have the meta mask extension installed can easily log in on the landing page by clicking a button. the user is routed to the official meta mask extension download page to install meta mask if it is not already installed. at window. ethereum, meta mask accesses a worldwide api into websites that its users browse. websites can access users' ethereum accounts with this api, read data from the associated blockchains, and recommend that users sign messages and transactions. the provider object's presence suggests an ethereum user. transitions framer motion an interactive design tool for websites and prototyping is called framer. building complete marketing sites, landing pages, online campaigns, and much more are its strong points. it covers each step of the design process, from interactive prototypes to graphic mock-ups, but its key advantage is publishing right from the canvas. because you can ship your design right away and all app transitions use it, framer is the quickest tool for building sites. type.js type of javascript programming language tool coin ranking api to integrate cryptocurrency prices into your app or website. gain access to high-quality data about all coins, like price history, circulating supplies, exchanges, trading pairs, and much more. the account page is customized to fetch current data about the user’s wallet coins or assets specifically (see fig. 5) backend development smart contracts, goerli benjamin kommey et al. design and implementation of a cloud based | 102 fig. 5. display of coin ranking api site to initiate testing, recharts were built on top of svg elements with a lightweight dependency on d3 submodules. a chart was customized by tweaking component properties and passing in custom components, quickly building the chart with decoupled, reusable react components. fig. 6 depicts the system dashboard with charts and recharts. fig. 6. display form functionality and validation were facilitated using the formik package which takes care of the repetitive functions—keeping track of values/errors/visited fields, orchestrating validation, and handling submission. by staying within the core react framework and away from magic, formik makes debugging, testing, and reasoning about forms intuitive. the system software application was deployed to the ipfs using fleek. fleek is a deployment platform employed in this system. fleek allows for continuous deployment in that, when any change is made and pushed to github, the change would automatically be seen on the deployed site. the fleek interface is as displayed in fig. 7. fig. 7. display of fleek interface smart contracts were written using solidity to define the logic for performing transactions and storing the data on the blockchain. contract application binary interface (abi) were developed from the smart contracts. this encodes the interface of the smart contract for the ui. that is, it tells the contract-abstraction library, for example ethers, what functions to provide. finally, a goerli wallet address was created and a goerli faucet was used to deposit virtual ether into the created wallet. goerli is a test net for deploying application into a sandbox, a test environment, for the development face of a project. alchemy was used to deploy the smart contract to the goerli test net. a network to deploy the contract to (either the ethereum main net or a test net) and wallet address to use was defined in the hardhat config file. the hardhat config file is a file that allows definition of conditions for deploying our smart contracts. lastly, 103 | international journal of informatics information system and computer engineering 3(1) (2022) 93-104 a context processor was defined in the frontend to provide the contract abi and to define other functions for manipulating the wallet. also, the processor monitors wallet related events, performed by user in the frontend. 4. results and discussion various tests were run to see if all functionalities were working as expected. that is, to be able to send and receive tokens on our decentralised application. the so called `goerli’ network was used as basis for intensive tests. this goerli network test was done before doing real testing on an ethereum network, since the deployment process is irreversible. the procedure for this testing is as described in the following. the goerli test net provided two users signed unto the platform via meta mask with virtual ethereum tokens through its goerli faucets. two users, user a and user b signed into the application via meta mask. with the virtual ethereum tokens provided by the goerli faucets, user a sent some tokens to user b’s account. this transaction was made possible to the meta mask wallet. the wallet address of user b was provided for user a. this address was used by user a in the locating user b’s account to the send the virtual ethereum tokens. the tokens were successfully received by user indicating a successful transaction. testing results were as expected, i.e., the system was able to successfully send and receive tokens seamlessly. 5. conclusion this paper has established the problems associated with centralised finance applications and successfully demonstrated the process of sending tokens from one ethereum based wallet to any other ethereum based wallet using meta mask. the project’s aim to develop a completely decentralised platform was achieved since the use of meta mask wouldn’t require any central body to make transactions and other functions. it was demonstrated that the frontend of the application could also be hosted on a decentralised platform and would allow for the exchange of tokens between wallets that are outside the ethereum network using bridging software. it would enable the direct deposit of fiat currencies as well and would integrate a platform to allow users to trade ethereum based tokens. references berengoltz, p., ofrat, i., & shaulov, m. (2021). u.s. patent application no. 17/172,794. binns, d. (2022). no free tickets: blockchain and the film industry. m/c journal, 25(2). buntinx, j. p. (2016). top 4 cryptocurrency projects created before bitcoin. digital currency business e-gold indicted for money laundering and illegal money transmitting: https://www.justice.gov/archive/opa/pr/2007/april/07_crm_301.html https://www.justice.gov/archive/opa/pr/2007/april/07_crm_301.html benjamin kommey et al. design and implementation of a cloud based | 104 fang, f., ventre, c., basios, m., kanthan, l., martinez-rego, d., wu, f., & li, l. (2022). cryptocurrency trading: a comprehensive survey. financial innovation, 8(1), 1-59. frankenfield, j. (2021). bitcoin. investopedia, feb, 18. griffith, k. (2014). a quick history of cryptocurrencies bbtc-before bitcoin. bitcoin magazine, 16. lam, c. h. (2019). u.s. patent application no. 16/429,075. lee, j., & cho, m. (2018). exeum: a decentralized financial platform for price-stable cryptocurrencies. arxiv preprint arxiv:1808.03482. mukhopadhyay, u., skjellum, a., hambolu, o., oakley, j., yu, l., & brooks, r. (2016, december). a brief survey of cryptocurrency systems. in 2016 14th annual conference on privacy, security and trust (pst) (pp. 745-752). ieee. nakamoto, s., & bitcoin, a. (2008). a peer-to-peer electronic cash system. bitcoin.– url: https://bitcoin. org/bitcoin. pdf, 4, 2. presthus, w., & o’malley, n. o. (2017). motivations and barriers for end-user adoption of bitcoin as digital currency. procedia computer science, 121, 8997. puspitawati, l., & ahmad, a. (2021). information system for forex investment and their effects on investment growth in foreign currencies. international journal of research and applied technology (injuratech), 1(1), 127-133. rahayu, s. (2022). implementation of blockchain in minimizing tax avoidance of cryptocurrency transaction in indonesia. international journal of research and applied technology (injuratech), 2(1), 30-43. toby shapshak: mobile money in africa reaches nearly $500bn during pandemic: https://www.forbes.com/sites/tobyshapshak/2021/05/19/mobilemoney-in-africa-reaches-nearly-500bn-during-andemic/?sh=1175069b3493 uriawan, w. (2020). swot analysis of lending platform from blockchain technology perspectives. international journal of informatics, information system and computer engineering (injiiscom), 1(1), 103-116. yusof, m. f., rasid, l. a., & masri, r. (2021). implementation of zakat payment platform for cryptocurrencies. azka international journal of zakat & social finance, 17-31. https://www.forbes.com/sites/tobyshapshak/2021/05/19/mobile-money-in-africa-reaches-nearly-500bn-during-andemic/?sh=1175069b3493 https://www.forbes.com/sites/tobyshapshak/2021/05/19/mobile-money-in-africa-reaches-nearly-500bn-during-andemic/?sh=1175069b3493 219 | international journal of informatics information system and computer engineering 3(2) (2022) 219-230 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 phishing website detection using several machine learning algorithms: a review paper alexander veach*, munther abualkibash school of information security and applied computing, eastern michigan university, ypsilanti, michigan, united states *corresponding email: aveach1@emich.edu 1. introduction phishing has become one of the most prevalent social engineering attacks in the digital environment. from personal accounts to corporate user accounts, all must be aware of the potential dangers of a phishing attack. this has led to an ongoing battle to prevent phishing attacks by blocking dangerous websites and communications. there are many methods to fight these attacks, with many looking to the new advancements in machine learning and artificial intelligence as a potential solution to phishing attacks. the method discussed in this paper is detecting phishing websites with machine learning algorithms. unfortunately, such a problem lacks a catch-all solution, which has led to the formation of multiple different approaches to the problem. for example, one solution could suggest designing a b s t r a c t s a r t i c l e i n f o phishing is one of the major web social engineering attacks. this has led to demand for a better way to predict and stop them in a commercial environment. this paper seeks to understand the research done in the field and analyse the next steps forward. this is done by focusing on what goes into the selection of proper features, from manual selection to the use of genetic algorithms such as adaboost and multiboost. then a look into the classifiers in use, neural networks and ensemble algorithms which were prominent alongside some novel approaches. this information is then processed into a framework for cloud-based and clientbased phishing website detection, alongside suggestions for possible future research and experiments that could help progress the field. article history: submitted/received 02 aug 2022 first revised 05 sept 2022 accepted 02 oct 2022 available online 20 oct 2022 publication date 01 dec 2022 aug 2018 __________________ keywords: artificial intelligence, data science, machine learning, phishing. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 219-230 https://doi.org/10.34010/injiiscom.v3i2.8805 alexander veach and munther abualkibash. phishing website detection using several machine…| 220 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 methods for on-hardware machine learning, which will limit the choice of algorithms to simpler versions but will allow for mass implementation. other solutions could focus on offloading the classification and model to a third-party service like microsoft azure or amazon web services, which circumvents the limitation of algorithms in exchange for another group of issues. including the differences in selecting features, where to gather the data, and much more there are a multitude of potential solutions with many looking for the most effective solution. the purpose of this paper is to look at the potential solutions and outline what the next steps for such research could be. 2. method to analyze the current popular solutions and implementation of anti-phishing technologies using machine learning and artificial intelligence a plethora of research was gathered from collage repositories and online journal sites such as jstor. once the multitude of research was gathered, which amounted to 91 papers. these 91 papers were then read and analyzed, taking the classifier and methods used into account and their differences. once that was complete, papers with relevance to the topic at hand and important for discussion were selected and used, the number of which is 14. 3. results and discussion 3.1. the material used the application of machine learning against phishing is not a new development and there has been a multitude of research done over the last few years. especially so for phishing urls. this is some of the relevant research that has come out in the last few years. there is sanchez-paniagua et al. (2022) who focused on analysing deep learning methods compared to other methods, namely ensemble and genetic selection algorithms. in their study they found that their model of using tf-idf + n-gram outperformed other methods by varying degrees. with the closest performers being within 0.5 points of accuracy while the weakest performers were behind as much as 10 points. the researchers also found that “...handcrafted url features decrease their performance over time, up to 10.42% accuracy in the case of the lightgbm algorithm from the year 2016 to 2020. for this reason, machine learning methods should be trained with recent urls to prevent substantial aging from the date of its release” (sanchez-paniagua et al., 2022). xiao et al. (2020) focused on using cnn with multi-head self-attention to determine if links were valid or phishing. by using mhsa, the researchers found better accuracy and speed compared to cnn-ltsm with a difference of 0.002 in cnn-mhsa’s favour. for future work, xiao et al. (2020) focuses on updating the model to take the html content into consideration to increase the accuracy further (xiao et al., 2020). a different direction was pursued by suleman and awan (2019), who focused on the use of generic algorithms such as “yet another generating genetic algorithm” or yagga. testing it against other gas found a 94.99% accuracy with an id3 classifier (suleman & awan, 2019). https://doi.org/10.34010/injiiscom.v3i2.8805 221 | international journal of informatics information system and computer engineering 3(2) (2022) 219-230 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 another example of study when it comes to genetic algorithms is subasi and kremic (2020) who compared adaboost and multiboosting when it came to testing phishing websites. the researchers found a high accuracy of 97.61% using an svm classifier with adaboost, however the cost of that accuracy is that svm adaboost reported a complexity, in seconds, of “8193.72” (subasi & kremic, 2020). another genetic algorithm study comes from alsariera et al. (2020) who focused on using their forest penalizing attributes algorithm that uses weight to deemphasize inconsequential variables. the team then compared the results to meta-learning variants of the algorithm specifically testing a bagging method and adaboost. of which they found that adaboosted forest penalizing attributes had an accuracy of 97%, beating the other accuracies of 96.26% for base classifier and 96.58% for bagged, and a speed where “...false alarm notifications are next to zero” (alsariera et al., 2020). a more unique approach is the one chen et al. (2020) took focusing on the visual similarity of websites to determine if it is a phishing website. it does this by using wavelet hashing and scale-invariant feature transform to determine similarity. the researchers found some success when using microsoft, dropbox, and bank of america as a comparison point, getting accuracy results of 98.14%, 98.61% and 99.95% respectively (chen et al., 2020). another unique approach is that of ali and malebary (2020), who used particle swarm optimization to improve detection of fraudulent phishing websites. by using the high speed pso model the team proposes feature weighting in much the same way a genetic algorithm operates. compared to the ga selection and weighting the team found that “...psobased feature weighting omitted between 7%-57% of irrelevant features” and found that classifiers using their method “...outperformed these machine learning models with applying ig, chi-square, wrapper, ga-based features selection, and ga-based features weighting” (ali & malebary, 2020). another approach takes the visual analysis of websites but then combines it with a neural network classifier. this approach is what abdelnabi et al. (2020) proposed which uses a triplicate network to compare websites to popular websites on alexa. by using the ensemble method with neural networks, they outline a potential future path for using website matching (abdelnabi et al., 2020). assefa and katarya (2022) focused on analysing other deep learning methods and their results and compared it to autoencoder, a form of unsupervised neural network. in the report they noted various limitations in other studies, noting issues such as non-comprehensive reports and compared their achievements to the autoencoder method. they found that autoencoder had an accuracy of 91.24% and that with better data mining techniques the performance could be improved (assefa & katarya, 2022). mandadi et al. (2022) focused on finding the most important features denoting three types, domain-based, html and javascript based, and address bar based features, with the total number of features under these three categories being considered was 17. once that was set, they tested the features with random forest and decision tree which gave values of 87.0% and 82.4% for accuracy respectively (mandadi et al., 2022). https://doi.org/10.34010/injiiscom.v3i2.8805 alexander veach and munther abualkibash. phishing website detection using several machine…| 222 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 saravanan and subramanian (2020) used ga feature selection alongside an artmap supervised neural network. artmap is made up of “a pair of selforganized adaptive resonance theory (art) modules arta and artb. these two modules are interconnected by an inter-art self-associative memory and an internal controller, whose objective is to maximize the predictive generalization and to minimize the predictive error. each art module is associated with f1 and f2 layers which act as a short term memory and a long term memory for category selection” (2020). this model also uses the firefly algorithm to determine which features are useful. the study found their own unique algorithm to be the best performing in all performance measures except for detection time, which svm performed better (saravanan & subramanian, 2020). mourtaji et al. (2017) also outlines which features they believe are best suited for detection. having five groups which are: lexical based analytics method, abnormal based feature, content-based analytics method, and an identity-based method. alongside these features they suggest a blacklist function on-top of these features. they used a linear regression classifier and reported an accuracy of 95.5% with a false positive rate of 1.4% (mourtaji et al., 2017). zhou and zhang (2022) propose a dualweight random forest algorithm that is “based on the combination of feature weight and decision tree weight”. the proposed classifier was then tested against random forest, random forest algorithm with decision tree weight, and dynamic random forest and had the highest accuracy with a value of 94.93% which was 2.22 points higher than the next highest which was dynamic random forest with 92.71 (zhou & zhang, 2022). 3.2. analysis phishing is one of the most dangerous and effective online fraud methods in existence today. this concern has led to the search for a so-called “silver bullet” that would protect potentially affected parties from phishing attacks. many have looked towards machine learning and artificial intelligence to create an application that, when used, would detect threats and adapt to them to create the ultimate defense. however there are many parts to consider including which classifier should be used for training, what attributes should be weighed to determine threat and which dataset is the best for training the model. the first major question is by which metric should such a model be trained around. should it be url focused, should it be based upon the content of the website itself, or should it be based on the websites meta content using tools such as whois. url based analysis is simple to implement and fast to process, but lacks other information from the website which can decrease accuracy. similarly analyzing the content of the web page alongside the url itself takes more time to execute for the benefit of more accurate results. some even suggest image recognition models such as chen et al. (2020) with their visual similarity model. then, when it comes to weighing features, some papers suggest using attribute selection algorithms such as adaboost, multiboost, or other genetic algorithms to predict which attributes lend themselves to correct identification such as suleman and awan (2019), subasi and kremic https://doi.org/10.34010/injiiscom.v3i2.8805 223 | international journal of informatics information system and computer engineering 3(2) (2022) 219-230 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 (2020), and alsariera et al. (2020). by using these machine proven attributes many hope to increase the efficiency of the used classification algorithms. subasi and kremic (2020) noted that “adaboost achieved the superior classification accuracy, with svm 97.61%” which beat their best accuracy single classifier result which was random forest which achieved “an accuracy of 97.26%”. another study done by sanchez-paniagua et al. (2022) reported that, when testing trained models based on data from 2016, 2017 and 2020: “...all models struggled to endure over time and their performance decreased when tested on the following years’ dataset” (sanchez-paniagua et al., 2022). thus showing the importance of an ever updating classification scheme. there are many offered solutions when it comes to what classifier to use, with two of the most common answers being neural network classifiers and random forest classification. random forest has been found by many researchers to be their choice of classifier in the studies surveyed. zhou et al. (2020) used a modified version of random forest, named double weighted random forest, and returned an accuracy 94.94% when using k-means clustering for feature selection. in studies that found other methods to be more effective such as sanchez-paniagua et al. the difference was only a 0.20 accuracy difference compared to lightgbm with 94.67 (sanchezpaniagua et al., 2022). however, some report a lower accuracy number, such as mandadi et al. (2022) who found a reported accuracy of 82.4% with 17 features using a phishtank dataset. this variance could be attributed to the differences in feature selection and the contents of the used datasets. another common solution is the use of neural network classifiers such as cnn, lstm, gnn and many others. neural network classification is recommended similarly to random forest with many studies finding high accuracy when predicting malicious phishing urls. as mentioned in the section prior, sanchezpaniagua et al. found that light bgm had the highest tested accuracy of the classifiers used with static feature selection on the piu-60k dataset (sanchezpaniagua et al., 2022). other studies have noticed similar results with other neural networks, specifically those with deep learning capabilities. xiao et al. (2020) applied multi-head self-attention, or mhsa, to a convolution neural network and found an accuracy rate of 0.9834 or 98.34 percent. the study proposed more solutions to increase that number even higher with their main worry being to “decrease the input of [url’s length parameter]” (xiao et al., 2020). novel application of the prior is also wellresearched. with a common focus on using visual detection, to detect pages that are too close to other pages as seen in abdelnabi et al’s work (2020). in their research they proposed a model that uses three convolutional models to determine phishing or not based on the similarity to other major pages collected from alexa. another unique approach is ali and malebary (2020) who propose a model based on particle swarm optimization feature weighing. which reportedly outperformed other weighting algorithms. like most topics there is not a singular silver bullet, so to speak, when it comes to predicting if a website is malicious or not. phishing methods commonly change to what is most efficient at that time which has led to a never ending conflict trying to https://doi.org/10.34010/injiiscom.v3i2.8805 alexander veach and munther abualkibash. phishing website detection using several machine…| 224 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 prevent said attacks. this has led to a focus on using genetic algorithms or other methods to create a curated list of features. as noted by sanchez-paniagua et al, “compared to machine learning algorithms, both cnn models obtained better results than handcrafted features” (sanchez-paniagua et al., 2022). by using deep learning models, a higher level of accuracy can be maintained at the cost of more costly requirements. neural network classifiers by design develop a much richer identification method, upon which they layer information in a way imitating human neurons, which requires more processing power than simple classifiers such as a decision tree classifier. these methods, when properly trained, can generate extremely accurate results. however, this in of itself is a much more costly method requiring a higher level of processing power commonly using high-end graphics cards designed for that explicit purpose such as the nvidia titan v. on the other end of the spectrum is random forest, or other ensemble classifiers, that instead rely on a series of classification tests to assure accuracy. thanks to this, ensemble classifiers require less processing power and have a better success rate with less data provided. however, random forest lacks the potential depth of learning that deep learning neural networks can possibly provide and is not adept when adapting to changes over time, as reported by sanchez-paniagua et al. (2022). then there are two further trains of thought when it comes to implementation, if the software should be designed to run off of the hardware it is installed upon or if the hardware should be run off of virtualized software through the cloud. both have their benefits and drawbacks, as offloading the processing better works when using devices such as mobile phones and other low powered devices. however, this builds a dependency on stable connection for the service to work, and a reliance on consistent service. this then creates specifications of an infrastructure that can support such needs. while using the physical machine itself limits the potential design of the model, as it must be customized to each device or be designed to work with most devices sacrificing customization. the benefit would be reliability, as the model would only require the model that is already trained and the processing power of the device executing it. this would limit potential downtime and other server connectivity issues, but could cost more in the long run for businesses implementing this method. another issue would be training the models in a reasonable way to adapt to changes in phishing techniques. something which sanchez-paniagua et al. (2022) found as much as a 10% decrease in accuracy as malicious phishing links change. the next most common solution was custom classifiers or unique analysis methods, or other similar methods, which made up nineteen of the ninety papers analyzed. these solutions focused on designing custom classifiers that would parse the target information, with claims that the unique solution was more effective than other common solutions. these classifiers are often similar to ensemble methods which combine classifiers in a multilayered approach. however, some are amalgamations designed to work as a single classifier instead of the normal multileveled https://doi.org/10.34010/injiiscom.v3i2.8805 225 | international journal of informatics information system and computer engineering 3(2) (2022) 219-230 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 classification that ensemble methods use which is why they have their own category. some of these solutions claim to have a success rate when tested of around 98 percent while others claim a much lower result. for example, saravanan and subramanian (2022) used a combination of a genetic algorithm to select important features and artmap, a neural network classifier based upon the firefly algorithm. there was also another group that had nineteen studies suggest its use. the deep learning methods are made up of such classifiers as cnn, dnn, gnn and their derivatives. these methods were used specifically to design evolving models that could potentially detect new attacks and adapt quickly. an issue with these studies is of course the resource intensive nature of deep learning methods. the method's resource intensive nature leaves only two options when it comes to potential implementation: require all hardware to meet the specification or offload the ai to a cloud-based solution. by requiring a dedicated gpu any company wishing to adopt will face a steep entry cost which will be a barrier to general adoption especially for major companies with tens of thousands of workers. the same is true for a cloud based solution as any corporation that wishes to adopt such a method will undoubtedly pay fees for such usage. something that was noticed in many of the reports is a lack of standardization when it comes to reporting the information gained from experimentation. several papers only reported the accuracy without any of the other data points leaving you to extrapolate how they reached that conclusion. this issue has been noted in other papers such as “intelligent phishing website detection using deep learning”, where assefa and katarya (2022) note that 3 of the papers analyzed failed to either provide enough details or the results reported were “not comprehensive”. this issue then compounds as a sizable group of papers would leave out important information such as the specifications of how they created their private dataset, and other key details needed to replicate their findings. this information is critical for understanding how efficient each method is. this can be remedied by having a standard for reporting the results of ai/ml for phishing detection. a solution would be to standardize what results are included in studies. this standard should require: a) the explicit location and name of which dataset was used, b) the algorithm used, c) explicit instructions on how the model was trained, d) an in-depth breakdown of false positives and negatives and true positives and negatives, and e) analysis execution speed. going forward there appears to be two paths when it comes to designing a defensive tool against fraudulent websites. the first approach would be focused on designing a client-focused service that would run a classifier on the hardware provided. the second approach would be to focus upon designing a cloudbased solution called through an api to offload the compute intensive work. both of these approaches have their own benefits and drawbacks, which will be discussed in greater detail in the next section, but either are a good beginning step for advancing anti-phishing measures. https://doi.org/10.34010/injiiscom.v3i2.8805 alexander veach and munther abualkibash. phishing website detection using several machine…| 226 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 3.3. example of a client-based solution the following is a proposed framework for a client-based solution for an antiphishing extension. the solution should be built in a browser native language, such as javascript, using the provided machine learning libraries such as tensorflow. when the website is accessed the extension will check a maintained whitelist which contains commonly used and trusted websites such as search engines, online office tools, and other trusted websites. then, if the website is not trusted the extension will harvest data needed for classification on the model used. for this example it will be assumed that an ensemble classifier such as random forest will be used. the classifier will account for multiple features including domain information, the url, and content on the website itself. something similar to the feature set suggested by mandadi et al. (2022), which lists dns record, website traffic, age of domain, end period of domain, iframe redirection, status bar customization, disabling right click, website forwarding, domain, ip address, “@” symbol, length, depth, redirection “//”, “http/https” in domain name, using url shortening services “tiny url”, prefix or suffix “-” in domain (mandadi et al., 2022). the extension should have a pre-built model based upon the above implemented in the extension, with updates to reflect trends in current phishing websites. while the extension classifies the website the extension should have an interim page that will update when classification is done to either send the user to the website or inform the user of the detected security risk. this model is considerably easy to implement and can theoretically be run on most modern workstations. this model also can be updated when performance drops due to changing trends in phishing to counteract the loss in accuracy, however doing so would require a consistent team to continuously watch the current trends in phishing websites. another weakness of this model is the potential for false positives and other accuracy issues, which would slow down the average user’s speed of use. the proposed model will also need to determine if the link is safe or unsafe rapidly, else earning the ire of the end user. these factors would need to be mitigated for a commercial implementation, by either optimizing the classification process, designing unique methods to obfuscate the methods in an unnoticed way, or other similar ideas (see fig. 1). https://doi.org/10.34010/injiiscom.v3i2.8805 227 | international journal of informatics information system and computer engineering 3(2) (2022) 219-230 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 fig. 1 a diagram of a simple client-based anti-phishing solution 3.4. example of a cloud-based solution where the prior solution is relatively simple to implement, the following is much more difficult due to the necessity of powerful computational processes, which are then hosted on either a public cloud service or a private cloud. this version would use a deep learning method, such as cnn-ltsm, which would be trained using information from repositories such as phishtank. the classifier should be guided to look at meta information, website content, and the website url itself. this trained model will then act upon information sent to it from client devices and determine if the site is a phishing website or a legitimate website. the model will then add that information into the next training set to continuously update the dataset to have it evolve naturally to counter new methods of phishing as they appear as suggested by sanchezpaniagua et al. (2022). this model, while simple to outline, is difficult to execute for practical use. for effective deep learning data needs to be consistently fed to the model for it to stay up-to-date. supporting this infrastructure would cost a lot of money or resources to execute effectively, alongside the customization needed to optimize the classification processes. ignoring those issues, another issue that one will run into is ensuring uptime for those dependent on the software. the cloud focused model requires consistent back and forth between all users and the classification service at all times for effective use. this also will require a lot of resources to implement. once the model is properly trained and maintained, it however has the potential for a higher accuracy than its ensemble based brother https://doi.org/10.34010/injiiscom.v3i2.8805 alexander veach and munther abualkibash. phishing website detection using several machine…| 228 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 above. in the deep learning studies surveyed for this paper, most reported an accuracy of 97% or more, trumping the average next highest classifier which was often the random forest algorithm. therefore, there is potential for cloudbased anti-phishing techniques powered by machine learning and artificial intelligence but the resource cost will limit effective implementation without serious capital investment (see fig. 2). 3.5. future work a prudent first step would be to standardize reporting of machine learning and artificial intelligence. currently there is no codified standard for reporting machine learning and artificial intelligence study results. some studies contain everything needed to replicate the experiments performed and how the conclusion was drawn; however other studies will leave out needed details for conclusive analysis or replication. mourtaji et al. (2017) for example outlines their own framework and show results from said framework without supplying the dataset used in testing, which they claim to have pulled from phishtank and alexa to populate. by providing the dataset used in testing to an online repository for verification it allows for doubt to be cleared and will be of great assistance to other researchers in the field. fig. 2 a diagram of a simple cloud-based anti-phishing solution https://doi.org/10.34010/injiiscom.v3i2.8805 229 | international journal of informatics information system and computer engineering 3(2) (2022) 219-230 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 by focusing on standards that ensure easy replication of results, and clarity within the information reported, other researchers will be able to work off of the research and develop the new technologies. therefore we would like to suggest a framework that would include these specifications for all reported testing: a repository containing the training dataset and testing dataset used, the features selected for classification, the classifier used alongside documentation of how to implement custom classifiers, the true positives and negatives alongside the false positives and negatives from resulting validation tests, precision rating, recall rating, accuracy rating, and f1 score. alongside this information there should be enough instruction for the reader to validate the paper by replicating the experiment within. by including this information it shall ensure reliable replication, which will make it easier to build upon thus helping the proliferation of information. on a more practical level the next step should be creating working models and testing them in live environments. by making a model, client or cloud based, will allow for researchers to see the practical shortcomings to these methods and correct them. once the shortcomings are known more development can take place evolving the field, which will help combat one of the most common threats on the internet. 4. conclusion phishing is one of the most common threats to cybersecurity in the current world. many organizations have become acutely aware of the potential danger of a successful attack. this has led to an increased focus on developing new technologies to prevent such attacks from taking place. by using machine learning and artificial intelligence many posit a learning defensive system that can prevent website phishing attacks and lower potential vectors for attack. currently there is no cure-all with many papers acknowledging the ever-changing nature of website based phishing attacks, preventing a permanent solution. however, a well automated system could go a long way to preventing websitebased phishing attacks and could be a useful solution for major organizations. most studies believe that a web extension for modern web browsers such as google chrome is where companies should look for future developments. a development of a working model for testing in live environments would do well in advancing the field by showing what potential shortcomings exist. finally, there is a lack of standardization in the reporting of data done in the multitude of studies focusing on the topic. to better advance the field in the focus of implementing anti-phishing ml/ai into working prototypes, a standard of reporting would make it easier to gather information. by always including the dataset used, the algorithm used, the instructions for training the model, a breakdown of the training and testing results and a record of time taken to execute a task, it would allow for information to be disseminated and processed faster which in turn could assist in the development of such antiphishing technologies. https://doi.org/10.34010/injiiscom.v3i2.8805 alexander veach and munther abualkibash. phishing website detection using several machine…| 230 doi: https://doi.org/10.34010/injiiscom.v3i2.8805 p-issn 2810-0670 e-issn 2775-5584 references abdelnabi, s., krombholz, k., & fritz, m. (2020, october). visualphishnet: zero-day phishing website detection by visual similarity. in proceedings of the 2020 acm sigsac conference on computer and communications security (1681-1698). ali, w., & malebary, s. (2020). particle swarm optimization-based feature weighting for improving intelligent phishing website detection. ieee access, 8, 116766116780. alsariera, y. a., elijah, a. v., & balogun, a. o. (2020). phishing website detection: forest by penalizing attributes algorithm and its enhanced variations. arabian journal for science and engineering, 45(12), 10459–10470. assefa, a., & katarya, r. (2022, march). intelligent phishing website detection using deep learning. in 2022 8th international conference on advanced computing and communication systems (icaccs) 1, 1741-1745. ieee. chen, j. l., ma, y. w., & huang, k. l. (2020). intelligent visual similarity-based phishing websites detection. symmetry, 12(10), 1681. mandadi, a., boppana, s., ravella, v., & kavitha, r. (2022, april). phishing website detection using machine learning. in 2022 ieee 7th international conference for convergence in technology (i2ct) (1-4). ieee. mourtaji, y., & bouhorma, m. (2017, october). perception of a new framework for detecting phishing web pages. in proceedings of the mediterranean symposium on smart city application (1-6). sánchez-paniagua, m., fernández, e. f., alegre, e., al-nabki, w., & gonzález-castro, v. (2022). phishing url detection: a real-case scenario through login urls. ieee access, 10, 42949-42960. saravanan, p., & subramanian, s. (2020). a framework for detecting phishing websites using ga based feature selection and artmap based website classification. procedia computer science, 171, 1083-1092. subasi, a., & kremic, e. (2020). comparison of adaboost with multiboosting for phishing website detection. procedia computer science, 168, 272-278. suleman, m. t., & awan, s. m. (2019). optimization of url-based phishing websites detection through genetic algorithms. automatic control and computer sciences, 53(4), 333-341. zhou, j., liu, y., xia, j., wang, z., & arik, s. (2020). resilient fault-tolerant antisynchronization for stochastic delayed reaction–diffusion neural networks with semi-markov jump parameters. neural networks, 125, 194-204. zhou, z., & zhang, c. (2022, may). phishing website identification based on double weight random forest. in 2022 3rd international conference on computer vision, image and deep learning & international conference on computer engineering and applications (cvidl & iccea) (263-266). ieee. https://doi.org/10.34010/injiiscom.v3i2.8805 hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 50 information and communication technology utilization for managerial planning among educational administrators in secondary schools in ilorin metropolis hammed olalekan bolaji1, tajudeen oluwafemi bolaji2* *1department of science education, faculty of education, al-hikmah university, ilorin-nigeria. 2department of educational management & counselling, faculty of education, al-hikmah *corresponding email: alikhlasschools@gmail.com a b s t r a c t s a r t i c l e i n f o this study examines how secondary school administrators in the ilorin metropolis use ict for managerial planning. the principal must effectively and efficiently organize, manage, and supervise the school's business for them to run successfully. this study's design is a survey-style descriptive study to determine how ict affects principals' administrative performance. the population of this study consisted of 292 participants from all 75 public secondary schools in the ilorin metropolis. the instrument for data collection was a structured questionnaire titled "information communication technology and principal's administrative effectiveness questionnaire" (ictpaeq) while reliability coefficient values of 0.85 were obtained. the data gathered from the field was evaluated using pertinent descriptive statistics such as percentages, mean, and standard deviation, while research question 4 was addressed using pearson production moment correlation (ppmc) statistics. this finding may be explained by the fact that most schools lacked these ict resources, which may have contributed to the finding. this finding revealed that secondary school principals do better administratively the more they use ict resources. this further implies that secondary school management effectiveness may be in danger in the absence of ict. this implied that ict is a crucial factor in the efficient administration of article history: received 18 dec 2022 revised 20 dec 2022 accepted 25 dec 2022 available online 26 dec 2022 aug 2018 __________________ keywords: ict, school administrator, administrative performance. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 50-64 51 | international journal of informatics information system and computer engineering 3(2) (2022) 50-64 1. introduction it is now generally acknowledged that information and communication technology (ict), which is driving globalization, deregulation, and innovation, are the main factors influencing the economic landscape. ict now serves as the foundation for economic growth that occurs at both the macro and micro levels, thus individuals who choose not to take part in such advances run the risk of becoming more marginalized by others. as a result, the use of ict and the management of knowledge has advanced learning across a variety of media, which has improved human understanding in the 21st century. ict is therefore a crucial component in maintaining managerial planning, teaching, and learning. more importantly, education should be highly valued because it can revolutionize any culture. it is the process of acquiring the information and abilities needed to maintain growth in all areas and sectors of life, for individuals, groups, and organizations. according to egwu (2016), a principal must effectively and efficiently organize, manage, and supervise the school's business for them to run successfully. as the head administrator of the secondary school, he is responsible for making optimal use of all available resources, including ict, to achieve school objectives. leaders at all levels must ensure that people and material resources are managed optimally for the educational system to accomplish its national plans and goals (nwune et al., 2016). nkwoh (2011) said that to effectively guide schools toward the achievement of educational goals, school principals must have a broad range of talents and be competent. according to carol and edward (2004), competency is the ability to complete a task successfully using knowledge, skills, attitude, and judgment. it is just the capacity and expertise needed to complete a specific task. the ability to successfully manage resources for productivity is referred to as managerial competency. the management of instructional programs, staff and student personnel administration, financial and physical resource management, and community relationship management are among the duties of school administrators as outlined by heller (2012). the formulation of school action plans and the viability of the institution depends heavily on the efficient management of both human and material resources. therefore, the teachers must have administrative abilities and demonstrate the to use and integrate technology effectively (teklemariam, 2009). the global revolution brought about by icts secondary schools. based on the findings of this study, the following recommendations were made for secondary school administrators to become more proficient and use ict technologies to enhance training for school administrators who are not yet ict proficient. hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 52 in teaching, learning, innovation, and institutional management must be used immediately (new media consortium, 2007). if the education industry is to fulfil its objectives and effectively compete on the global stage, school managers play a crucial role in the usage and integration of ict. ict adoption in schools won't be successful unless the principal fully supports it, acquires the necessary knowledge, and offers staff professional development and encouragement during the transition (wilmore, 2000). 1.1. statement of the problem stakeholders in the ilorin metropolitan have expressed worry over the management situation in secondary schools. this worry has focused on the school principals' failure to successfully perform their administrative tasks as a result of the significant growth in student enrollment, class size, pupils, and teacher data, which is typical of nigerian secondary schools (okon et al., 2015). administrative duties in schools would inevitably be processed using advanced tools and infrastructure, like ict. what is more important to worry about is whether or not there are adequate information and communication technology (ict) facilities for effective secondary school administration, and if there are, how well they are being used. this is not unrelated to the government of nigeria's inadequate funding for the education sector, secondary school administrators' inability to keep up with the rate of ict development, and the lack of personnel with the necessary skills to manage ict at all levels of administration. (muchiri, 2014), the ineffective institutional policies that encourage and regulate the use of ict, and the intermittent or nonexistent electrical supply in schools. observations made in our secondary schools indicated that this deficiency can have a significant impact on the managerial efficiency of secondary school principals. this raises the question of how much the availability and use of ict facilities affect the administrative effectiveness of public secondary school principals in the ilorin metropolis. 1.2. purpose of the study this study's primary goal is to examine how secondary school administrators in the ilorin metropolis use ict for managerial planning. in particular, the study looked into: 1. assess the extent to which secondary schools in the ilorin metropolitan had ict facilities. 2. examined how ict was used by secondary school principals in the ilorin metropolis for administrative purposes. 3. determined the degree of performance of secondary school principals in carrying out their regular administrative duties in the ilorin metropolis 4. analyze the association between secondary school principals' administrative effectiveness and the use of ict facilities in the ilorin metropolis. 1.3. research questions the following research questions were raised: 53 | international journal of informatics information system and computer engineering 3(2) (2022) 50-64 1. what degree of ict infrastructure is present in secondary schools in the ilorin metropolis? 2. what percentage of secondary school principals in the ilorin metropolis use ict resources for administrative tasks? 3. what level of efficiency do the ilorin metropolis' secondary school principals exhibit when carrying out their regular administrative duties? 4. does the use of ict facilities in the ilorin metropolitan significantly affect the administrative effectiveness of secondary school principals? 2. literature review conceptualization of information and communication technology according to reports, the word information and communications technology (ict) was first used to refer to computers' capabilities for communication around the beginning of the 1990s, replacing the term information technology (it). information and communication technology (ict) is the fusion of communication, information, and technologies that are based on common digital technology, according to egoeze, misra, akman, and colomopalacios (2014). ict is thus defined as an electronic instrument, equipment, or device that is used to gather, process, store, retrieve, or transfer information. according to salisu (2014), information and communication technologies (icts) include telecom networks as well as related hardware, software, and services (salisu, 2014; shah, 2014). ict was divided into four (4) different categories by: communication technologies, network technologies, computer technologies, and mobile technologies. all forms of media used for transferring voice, video, data, or multimedia are included in communication technologies. personal area networks (pan), campus area networks (can), intranets, extranets, local area networks (lans), wide area networks (wans), and the internet are examples of network technology. computer technologies include all types of removable storage devices, including optical discs, disks, flash memory, video books, multimedia projectors, interactive electronic displays, and modern personal computers that are constantly being developed (pcs). e-learning also makes use of mobile technology like mobile phones, personal digital assistants (pdas), palmtops, etc. that contain information as their physical object. ict was further divided into five groups by (brinda et al. 2016; ereh et al., 2012), including computing facilities and services; film/tape-based facilities, such as slide projectors, microfiche readers, micro card readers, and microprint readers; reproduction facilities, such as photocopying and duplicating machines; and telecommunication facilities. the degree to which a secondary school can offer all the ict resources required is a reflection of the prestige of the institution. ict use in nigerian education is hindered by several variables. these include a lack of finance to support the technology's acquisition, a lack of training, a lack of staff enthusiasm to utilize icts as teaching tools in the classroom, and more (oluwalola, 2017; oyedeji, 2015). arikewuyo (2009) added that the principal's duties include, but are not limited to: managing and allocating hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 54 school resources effectively; allocating school space effectively; ensuring satisfactory standards of upkeep and cleanliness of school facilities; planning staff development; and directing curriculum change and implementation (arikewuyi, 2009). according to adeniyi and omoteso (2014), in carrying out these duties, principals must show they are capable of leading by showcasing their professional expertise, organizational and managerial skills, and capacity to develop and implement sound school policies (adeniyi & omoteso, 2014; adesina, 2015). an international notion of utmost significance is administration. adeniyi and omoteso (2014) define administration as the process of coordinating people and resources within a company to achieve predetermined goals. additionally, the term "administration" refers to tasks like carrying out organizational objectives and choices as well as doing in-depth research on these tasks. performance is a difficult concept to define and comprehend. effectiveness is frequently used as a synonym for other words including effectiveness, efficiency, productivity, and competency. according to okon et al. (2011), performance is the degree of accomplishment given the time, resources, and circumstances. as a result, performance refers to a principal's execution of statutory tasks or functions that are directed at achieving goals established by the school (okon et al., 2011). accordingly, achieving daily management goals that are both feasible and of the utmost importance to the school constitutes efficient administration. as a result, a proactive secondary school principal's duties include organizing the available material and human resources and applying them consistently to the fulfilment of educational goals (national open university nigeria, 2014). the methodical coordination of human, material, and financial resources for the resolution of ongoing administrative issues to realize the planned objectives of secondary education is the principal's administrative performance. studies on ict and principal’s administrative performance in recent years, several surveys have been developed to gather data on the degree to which schools are building their capacity to integrate ict into learning, teaching, and management operations. oboegbulem and ugwu (2013) examined 30 schools in the southeastern states, including abia, anambra, ebonyi, enugu, and imo, which had access to ict and the internet (oboegbulem & ugwu, 2013). the results demonstrated, among other things, the need of using ict in school administration. especially during the current globalization period. however, due to school administrators' inability to manage ict equipment in schools properly, secondary schools only use ict to a very limited extent. ex-post facto research (okon et al., 2015) among 255 principals from 85 public secondary schools in akwa ibom state, nigeria, found a substantial correlation between the use of ict for record-keeping and communication, as well as the administrative performance of principals. 55 | international journal of informatics information system and computer engineering 3(2) (2022) 50-64 results from a related study conducted by oyedemi (2015) on the opinions of 120 administrators regarding the use of ict for efficient school management in the ilesa local government area of osun state showed that school principals have a favourable reaction to the usage of ict tools (oyedemi, 2015). according to subair and bada (2014), who conducted a similar study in osun state with 100 principals of public secondary schools, school administrators are aware of the value of ict in school administration, but the main challenge is a lack of the necessary skill and knowledge to use these resources. additionally, it was discovered that the majority of schools lack the necessary ict resources and those very few principals can use the ict tools at their disposal for administrative functions. the study also discovered that the majority of principals used print technologies for a variety of administrative tasks. in nigeria's akwa ibom state, (etudor-eyo, et al. 2012) researched 396 secondary school administrators' use of ict and communication efficacy. the results showed that administrators' use of ict and their effectiveness in communication are both high; there is a significant positive relationship between administrators' use of ict and their effectiveness in communication, and the use of ict significantly predicts secondary school administrators' communication effectiveness. a little body of knowledge about ict and administrative performance in the research area (ilorin metropolis) and the geopolitical region (north-central geopolitical zone), both of which are located in nigeria, is known as a result of the aforementioned empirical review. the topic coverage of these studies is one of the primary constraints of current research, which justifies the necessity for additional study. additionally, the literature review reveals that despite a large number of studies on this research project, the findings do not demonstrate any consistency in the relationship between the variables under examination. these findings point to a glaring vacuum in the literature that this current study aims to remedy. 3. methodology this study's design is a survey-style descriptive study to determine how ict affects principals' administrative performance. according to atunde (2011), descriptive research allows the researcher the chance to sample the opinions of a sizable number of samples from the study population to draw conclusions and make generalizations based on the replies received (atunde, 2011). the study area is the metropolis of ilorin. principals and vice-principals of secondary schools in the ilorin metropolis participated in the survey. ilorin west, ilorin south, and ilorin east are the three local government areas (lgas) that make up the metropolis. there are 75 public secondary schools in the ilorin metropolis overall (kwara state ministry of education and human capital, 2018; the federal republic of nigeria, 2013). the population of this study consisted of 292 participants from all 75 public secondary schools in the ilorin metropolis, including 75 principals and 217 vice-principals. 45 of the 75 public secondary schools in ilorin metropolis were sampled using the hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 56 stratified random sampling technique (15 schools were chosen from each of the lgas). additionally, from each of the 45 public secondary schools used for this study, the principal and two viceprincipals were chosen. this increased the number of responders who were used to 135 in total. the instrument for data collection was a structured, validated, and pretested questionnaire titled "information communication technology and principal's administrative effectiveness questionnaire" (ictpaeq) (reliability coefficient values of 0.85, 0.74, and 0.71 were obtained for sections b, c, and d, respectively). there are two sections to the questionnaire: parts a and b. the questions in part a ask respondents for personal information. part b is divided into four sections (a to d). the 15 items in section a were used to determine if there was ict equipment in the schools. ten questions make up section b, which is meant to gather data on how extensively principals use ict. questions in section c ask respondents to provide information on how successfully they use ict tools to carry out their administrative responsibilities. items in section d are included to help you learn more about the barriers to effective ict use in secondary schools. the questionnaire has parts a and b with options such as "available" and "not available," "frequently used," "occasionally used," "seldom used," and "not used," and sections c and d with options such as "strongly agree," "agree," "disagree," and "strongly disagree." the questionnaire was given to the subjects by the researcher with the assistance of three research assistants. the straightforward method allowed for prompt completion and return of the questionnaire copies. 96.3 per cent of those who completed the survey were responsive (that is 130 out of 135 administered questionnaires were returned and filled correctly). to address research questions 1, 2, and 3 respectively, data gathered from the field was evaluated using pertinent descriptive statistics such as percentages, mean, and standard deviation, while research question 4 was addressed using pearson production moment correlation (ppmc) statistics. research question 1: what degree of ict infrastructure is present in secondary schools in the ilorin metropolis? when interpreting respondents' responses to questions about availability, a response with an overall percentage score between 75 and 100 per cent is considered high, a response with an overall percentage score between 50 and 74 per cent is considered moderate, and a response with an overall percentage score of less than 50 per cent is considered low. according to each of the specific research topics, the findings of this study were displayed in table 1. 57 | international journal of informatics information system and computer engineering 3(2) (2022) 50-64 table 1. the level of availability of ict facilities in secondary schools in the ilorin metropolis s/n items available n = 130 f & % not available f & % 1. computer (desktop) 84(64.6%) 46(35.4%) 2. printer 78(60.0%) 52(40.0%) 3 internet services 72(55.4%) 58(44.6%) 4 projector 24(18.5%) 106(81.5%) 5 projector screen 24(18.5%) 106(81.5%) 6 photocopying devices/xerox machines 70(53.8%) 60(46.2%) 7 scanning machine 48(36.9%) 82(63.1%) 8 computer accessories 79(60.8%) 51(39.2%) 9 software 33(25.4%) 97(74.4%) 10 radio 104(80.0%) 26(20.0%) 11 television 37(28.5%) 93(71.5%) 12 satellite disc 31(23.8%) 99(76.2%) 13 handset/mobile phone 130(100.0%) 14 laptop 38(29.2%) 92(70.8%) 15 fax machine 130(100.0%) overall 43.7% 56.3% table 1 shows that the sampled secondary schools had a poor availability of information and communication technology (ict) facilities (43.7 per cent). the majority of secondary schools lacked the majority of ict resources. the shortfall ranged from 0% for a fax machine to 18.5% for a projector and projector screens, 25.4% for software, and 23.8% for a satellite disc. 35.4 per cent of secondary schools lacked a desktop computer set, which ought to be a standard component of all administrative offices. however, the cell phone (130 or 100.0%) and radio (80.0%) were the most accessible ict resources in public secondary schools in ilorin city. research question 2: examined how ict was used by secondary school principals hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 58 in the ilorin metropolis for administrative purposes. analysis of research question 2 made use of mean computation. this led to the classification of a mean score of above 2.5 as an acceptable response and a mean score of below 2.5 as a non-accepted response (rejected). the following was the interpretation of the individual and grand mean scores: high extent (he) was defined as 3.25 to 4.00, moderate extent (me) as 2.50 to 3.24, and low extent (le) as mean values below 2.50. (le) (table 2). table 2. the level of ict used by secondary school principals in the city of ilorin for administrative purposes s/ n items n = 130 mean s. d decision 16 a device that processes and stores data for efficient management 3.30 0.78 he 17 equipment for printing documents 2.75 0.75 me 18 internet browsing services 2.64 0.73 me 19 handset for communicating the most recent information about school-related concerns to staff, parents, and students. 3.96 0.94 he 20 scanners for scanning documents, such as passports. 2.19 1.12 le 21 radio for monitoring the most recent and relevant events or information worldwide 3.50 0.85 he 22 flash drives are a type of computer accessory used for data and information storage. 2.89 0.78 me 23 satellite disc for remote viewing of international programming 2.46 1.03 le 24 reproduction of staff, student, and school documents using a photocopier 3.27 0.85 he 25 for writing and designing, use programs like microsoft office and corel draw. 1.98 1.07 le overall mean 28.95 8.90 grand mean 2.90 0.89 moderate extent 59 | international journal of informatics information system and computer engineering 3(2) (2022) 50-64 the amount of ict used for administrative purposes by secondary school principals in the ilorin metropolitan is shown in table 2. the table demonstrates that the mean scores on items 19, 21, 16, and 24 are respectively 3.96, 3.50, 3.30, and 3.27, indicating that principals use their cellphones to communicate with staff, students, and other principals as well as with students and staff; radios to keep up with global news; computers to type, process, and store data for effective management; and photocopiers to make copies of documents. additionally, the mean values on items 22, 17, and 18 of 2.89, 2.75, and 2.64, respectively, show that principals use computers and computer accessories to process data, printers to print documents, and internet services to browse. however, a limited percentage of exindicatecipals employ software of various types, scanners for scanning passports, documents, etc., and satellite discs for viewing foreign programs remotely, as indicated by mean values of 2.46, 2.19, and 1.98 on items 23, 20, and 25, respectively. as a result, the grand mean value of 2.90 indicated modest ict use for administrative purposes by principals in public secondary schools in the ilorin metropolitan. research question 3: what level of efficiency do the ilorin metropolis' secondary school principals exhibit when carrying out their regular administrative duties? mean computation was used to analyze research question 3. accordingly, a mean score of above 2.5 was considered an appropriate response, whereas a mean score below 2.5 was considered an unacceptable response (rejected). the scores, both individual and overall, were interpreted as follows: highly level (hl) values ranged from 3.25 to 4.00, moderately level (ml) values ranged from 2.50 to 3.24, and low level (ll) values ranged from mean values below 2.50 (table 3). table 3. the level of principals' performance in carrying out their standard administrative responsibilities in secondary schools in the city of ilorin s/n items n = 130 mean s. d decision 26 preserving data that can be updated on personnel or student data. 2.60 0.81 ml 27 composing and sending mail 3.09 0.60 ml 28 preserving an updated inventory of the school's assets 3.29 0.74 hl 29 keeping an updated inventory of the school's assets in storage 1.75 1.04 ll hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 60 30 design and printing of student evaluations and testimonies. 2.42 0.91 ll 31 admission of new students and registration of current ones. 2.95 0.69 ml 32 online registration for public exams is available for students. 2.64 0.71 ml 33 early budget planning for the school. 2.56 0.73 ml 34 keeping and accessing student disciplinary records. 2.80 0.82 ml 35 creating internal school memos 2.63 0.78 ml 36 preparing the workload of teachers 2.75 0.75 ml 37 creating a school schedule 2.70 0.81 ml 38 spreading information both inside and outside of the school. 2.85 0.78 ml 39 compiling academic performance data for students 2.66 0 .77 ml 40 keeping reliable records of students' academic progress. 2.24 1.02 ll overall mean 39.79 11.26 grand mean 2.65 0.75 ml table 3 results showed that the grand mean value of 2.65 is higher than the 2.50 criterion limitations. this suggests that secondary school principals performed their usual administrative duties at a somewhat high level. research question 4: does the use of ict facilities in the ilorin metropolitan significantly affect the administrative effectiveness of secondary school principals? (table 4). table 4. the use of ict resources and administrative effectiveness are correlated descriptive statistics mean std deviation n usage of ict 28.9523 8.90304 130 administrative performance 39.7853 11.25708 130 61 | international journal of informatics information system and computer engineering 3(2) (2022) 50-64 correlations usage of ict administrative performance usage of ict pearson correlation 1 631 sig. (2-tailed) 000 n 130 130 administrative performance pearson correlation 631 1 sig. (2-tailed) 000 n 130 130 correlation is significant at the 0.05 level (2-tailed). as seen in table 4, the calculated r-value (0.631) is greater than the critical r-table value (0.195) at 0.05 significance levels for 128 degrees of freedom. hence, the null hypothesis is rejected. this demonstrates that there was a strong positive correlation between secondary school principals' administrative effectiveness in the ilorin metropolis and their use of ict facilities. 4. discussion of findings according to the study's findings, there were few information, communication, and technology (ict) facilities available in public secondary schools in the ilorin metropolitan. this result was in line with a previous study by adeyemi and olaleye (2010), who discovered that secondary schools in ekiti state had limited access to ict equipment (adeyemi & olaleye, 2010; anukam et al., 2012). this finding is in line with that of subair and bada (2014), who claimed that only a small percentage of osun state's public secondary schools have the necessary ict resources, making it difficult for principals to use those resources for administrative functions. results from the second research question showed that secondary school principals in the ilorin metropolitan used ict to a modest degree for administrative purposes (grand mean value of 2.90). this finding may be explained by the fact that most schools lacked these ict resources, which may have contributed to the finding. the conclusions of this study are refuted by adeyemi and olaleye's (2010) study, which found that secondary school principals in ekiti state used ict equipment at a low rate. the current study also found that principals performed only moderately in carrying out their designated administrative obligations. hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 62 this finding might not have been unrelated to the issue of a lack of ict resources in the schools as well as their sparing use of those resources. this result, however, contradicts that of adeyemi and olaleye (2010) who noted that the level of management of secondary schools in ekiti state, nigeria, was low. it is therefore consistent with other researchers' findings (makewa et al., 2013). further research revealed a substantial correlation between the use of information and communication technology (ict) resources and secondary school administrators' administrative performance. this revealed that secondary school principals do better administratively the more they use ict resources. this further implies that secondary school management effectiveness may be in danger in the absence of ict. this result is consistent with that of oboegbulem and ugwu (2013), who reported that the use of ict in school administration is essential, particularly in this era of globalization, but that the extent of their application in secondary schools is very low because school administrators in the south eastern states of nigeria lack the skills to manage ict facilities for the efficient administration of schools. this result was in line with that of etudor-eyo et al. (2012), who claimed that ict had a significant impact on the efficient operation of secondary schools. 5. conclusion according to the study's findings, information and communication technology (ict) significantly affects secondary school principals' administrative performance in the city of ilorin. this implied that ict is a crucial factor in the efficient administration of secondary schools. however, the researcher concluded that secondary schools in nigeria are not yet prepared for technological progress because of the moderate availability and usage of ict facilities. based on the findings of this study, the following recommendations were made; the government should make ict technologies accessible in all secondary schools so that administrators may become more skilled and use them. governments and ngos (nongovernmental organizations) may on occasion provide school administrators with improved training if they are not yet ict-savvy. for school administrators who are not yet ict-savvy, governments and ngos (non-governmental organizations) may on occasion offer improved training. references adeniyi, w. o., & omoteso, b. a. (2014). emotional intelligence and administrative effectiveness of secondary school principals in southwestern nigeria. international journal of psychology and behavioral sciences, 4(2), 79-85. adesina, m. o. (2015). ict: its relevance in the teaching and learning of physical education in nigeria. journal of emerging trends in educational research and policy studies, 6(3), 236-239. 63 | international journal of informatics information system and computer engineering 3(2) (2022) 50-64 adeyemi, t. o., & olaleye, f. o. (2010). information communication and technology (ict) for the effective management of secondary schools for sustainable development in ekiti state, nigeria. american-eurasian journal of scientific research, 5(2), 106-113. angie, o., & ugwu, r. n. (2013). the place of ict (information and communication technology) in the administration of secondary schools in south eastern states of nigeria. anukam, i. l., okunamiri, p. o., & ogbonna, r. n. o. (2012). basic text on educational management. arikewuyo, m. o. (2009). professional training of secondary school principals in nigeria: a neglected area in the educational system. florida journal of educational administration & policy, 2(2), 73-84. atunde, m. o. (2011). influence of management information system on academic staff effectiveness in kwara state colleges of education. unpublished m. ed thesis, national open university of nigeria. brinda, t., mavengere, n., haukijärvi, i., lewin, c., & passey, d. (2016). stakeholders and information technology in education. springer international publishing. egoeze, f., misra, s., akman, i., & colomo-palacios, r. (2014). an evaluation of ict infrastructure and application in nigeria universities. acta polytechnica hungarica, 11(9), 115-129. ereh, c. e., & okon, n. n. (2015). keeping of teachers’ record and principals’ administrative effectiveness in akwaibom state secondary schools, nigeria. international journal of education, learning and development, 4(1), 4044. etudor-eyo, e., emah, i. e., & ante, h. a. (2012). the use of ict and communication effectiveness among secondary school administrators. educare, 4(2). makewa, l., meremo, j., role, e., & role, j. (2013). ict in secondary school administration in rural southern kenya: an educator’s eye on its importance and use. international journal of education and development using ict, 9(2). muchiri, g. m. (2014). factors influencing school principals’ integration of ict in administration of public secondary schools in githunguri sub county, kiambu county, kenya (doctoral dissertation, university of nairobi). national open university nigeria (2014). course material on theories and practice of public administration (pad813). national open university of nigeria press. okon, f. i., akpan, e. o., & ukpong, o. u. (2011). financial control measures and enhancement of principals’ administrative effectiveness in secondary schools in akwa ibom state. african journal of scientific research, 7(1), 335342. okon, j. e., ekaette, s. o., & ameh, e. (2015). information and communication technology (ict) utilization and principals' administrative effectiveness in public secondary schools in akwa ibom state, nigeria. african educational research journal, 3(2), 131-135. hammed olalekan bolaji and tajudeen oluwafemi bolaji. information and communication …| 64 oluwalola, f. k. (2017). record keeping, information and communication technology in school management. ed. olubor, r.o. et al. educational management: new perspectives. amfitop books. oyedeji, n.b (2012). management in education: principles and practice (revised edition). success educational services. oyedemi, o. a. (2015, july). ict and effective school management: administrators’ perspective. in proceedings of the world congress on engineering 1, 1-3. salisu, r. o. (2014). information and communication technology (ict) and registrars’ administrative effectiveness in kwara state colleges of education. shah, m. (2014). impact of management information systems (mis) on school administration: what the literature says. procedia-social and behavioral sciences, 116, 2799-2804. the federal republic of nigeria (2013). national policy on education. nerdc press. moksud alam mallik. an efficient fuzzy clustering algorithm for mining user session ...| 80 an efficient fuzzy clustering algorithm for mining user session clusters on web log data moksud alam mallik1,2*, nurul fariza zulkurnain1 1international islamic university malaysia, kuala lumpur, malaysia. 2vnr vignana jyothi institute of engineering & technology, hyderabad, india. *corresponding email: 1alammallik_m@vnrvjiet.in a b s t r a c t s a r t i c l e i n f o data mining is extremely vital to get important information from the web. additionally, web usage mining (wum) is essential for companies. wum permits organizations to create rich information related to the eventual fate of their commercial capacity. the utilization of data that is assembled by web usage mining gives the organizations the capacity to deliver results more compelling to their organizations and expanding of sales. client access patterns can be mined from web access log information using web usage mining (wum) techniques. because there are so many end-user sessions and url resources, the size of web user session data is enormous. human communications and non-deterministic browsing patterns increment equivocalness and dubiousness of client session information. the fuzzy set-based approach can solve most of the challenges listed above. this paper proposes an efficient fuzzy clustering algorithm for mining client session clusters from web access log information to find the groups of client profiles. in addition, the methodologies to preprocess the net log data as well as data cleanup client identification and session identification are going to be mentioned. this incorporates the strategy to do include choice (or dimensionality decrease) and meeting weight task assignments. article history: received 18 dec 2021 revised 20 dec 2021 accepted 25 dec 2021 available online 26 dec 2021 aug 2018 __________________ keywords: data mining, web usage mining (wum), data preprocessing, fuzzy clustering. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(2) (2021) 80-93 81 | international journal of informatics information system and computer engineering 2(2) (2021) 80-93 1. introduction data mining, the extraction of hid judicious information from immense informational collections, is a staggering new development with the phenomenal potential to help associations revolve around the fundamental information in their data stockrooms. information mining instruments anticipate future examples and work on them, allowing associations to make proactive datadriven decisions. utilizing a blend of ai, measurable investigation, demonstrating methods, and data set innovation, information mining discovers designs and unobtrusive connections in information and construes decisions that permit the forecast of future outcomes. data mining (information disclosure from information) is the extraction of fascinating for example non-immaterial, verifiable, ahead-of-time dark, and conceivably important examples or information from a huge proportion of information. it changes locally very well and may be alluded to as information revelation (mining) in data sets (kdd), information, extraction, information, design investigation, and so forth (han et al., 2012; zahid et al., 2011; cooley et al., 1997). web mining is defined as the disclosure and evaluation of useful data from the world wide web in a broad sense. there are two sections to web mining: web content mining and web utilization mining are two types of web mining. web use mining is the automated disclosure of user access patterns from web servers. every business collects a significant amount of data on a daily basis in its operations. web servers generate this information, which is saved in server access logs. examining server access log data helps the organization to focus on lifetime estimation of customers, showcasing strategies for products, effective promotional campaigns, etc. it also helps in rebuilding websites to represent the organization and promote their products and services in a better way in www. web mining is by and large isolated into two parts. the first part is secondary in space; it converts web data into an appropriate exchange structure. this combines exchange id preparation and information inclusion. the subsequent part is space selfsufficient applications like general information mining and example coordinating with methods like clustering (cooley et al., 1997). preprocessing, information extraction, and examination outcomes are all included in wum. the preprocessing stage of web-use mining aims to convert unprocessed web log data into a large number of customer profiles. each of these profiles receives a plan or a number of urls related to a customer session. the preprocessing stage in web-use mining changes the harsh snap stream data into a gettogether of customer profiles. each of these forms contains a set of urls that correspond to a client session. for different preprocessing activities, such as data fusion and cleaning, user and session identification, and so on, several algorithms and heuristic methods are used. convergence of log files from several web servers is referred to as data fusion. data cleaning incorporates assignments, for example, eliminating unnecessary references to inserted objects, style documents, illustrations, or sound records, and disposing of references because of bug routes. by doing away with an undesirable substance like this we can lessen the size of the input file and make the mining moksud alam mallik. an efficient fuzzy clustering algorithm for mining user session ...| 82 errand efficient. so, during preprocessing we will clean the data, identify the user by using the ip address and identify the user session by using time-oriented heuristics. we can assign weight to urls based on the number of times they are accessed in different sessions also weight can be assigned to a session according to the number of urls present in it (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). when user sessions are found we can utilize them for clustering. little sessions will be removed because it shows disturbances in the data. rather than straightforwardly removing it, we can utilize a fuzzy set_theoretic way to deal with it. direct elimination of minimal estimated sessions may achieve a loss of a gigantic proportion of data. so, we can relegate weight to all sessions considering the number of urls got to by the session (see figure 1). figure 1. structure of web usage mining after this, we can apply the fuzzy clustering algorithm to recognize user session clusters. fuzzy membership is promoted by fuzzy clustering. in this case, a single informational index can be used by many groups. it suggests that one informational collection can find a place with a few bunches all the while. every informational index will have a degree of enrollment in each group; some groups will have a high level of participation, while others will have a low level of enrollment. the value of participation will range from zero to one. the total assessment of the participation of one meeting to each bunch of habitats will be one. data fuzzy clustering ought to oversee fit reality. for instance, if an informational index is on the limit between at least two bunches fluffy grouping will give it halfway participation among bunches (bezdek et al., 1984). in fuzzy clustering, each datum point has relegated participation worth to every one of the clusters. if the membership value is zero the data is not a piece of that cluster. no zero value shows that the data is attached to that cluster. membership value will be always between zero and one. here we can discover similar user access patterns i. e. same url patterns by applying the fuzzy clustering algorithm. the output of this step will be separate user session clusters it (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). the literature review is found in section 2, the proposed algorithm is found in section 3, the test results are found in section 4, and the conclusion and future improvements to this research are found in section 5. 2. literature survey digitized information is easy to capture and storing it is very cheap. so gigantic measure of data has been put 83 | international journal of informatics information system and computer engineering 2(2) (2021) 80-93 away in distinctive sorts of databases and other types of storage. the data storage frequency is developing at an exceptional rate. this developing data is amassed in various huge data storages. this sort of circumstance requires intense apparatuses to grasp knowledge from this ocean of information. with the exceptionally high development of data sources open on the world wide web, it has wound up continuously indispensable for clients to use customized instruments in finding the needed information resources, and to follow and dissect their utilization designs. so, there is a necessity to create server-side and client-side tools that mine knowledge adequately (cooley et al., 1997). web usage mining is the revelation of client access designs from web servers. how clients are getting to a webpage is critical to building the use of the site by clients. there are three steps to it. preprocessing, pattern extraction, and examination of the results. different forms of sounds are removed during the preprocessing stage. the user and session identification process will be completed in this stage. a wide variety of pattern extraction techniques are available like clustering, path analysis, etc based on the needs of the analyst. once web usage patterns are discovered there are different types of techniques and tools to analyze and understand them. a gigantic amount of unessential data is available in input web access logs. many user sessions and url resources makes the dimension of web-user session data very high. human interactions and nondeterministic browsing patterns increase the ambiguity and vagueness of user session data. the world wide web is a massive, dynamic data source that is both architecturally complex and constantly evolving. as a result, it is a fertile ground for data mining and web mining. using various information mining methodologies, web mining can be utilized to extract valuable information from the internet. the majority of web information is unlabeled, dispersed, heterogeneous, semicoordinated, time-moving, and multidimensional. the following categories of data can be found on the internet: (i) the substance of real web pages (ii) intra-page constructions of the website pages. (iii) inter-page structures decide linkage structures between website pages. (iv) we use information depicting web (v) user profiles incorporate demographic and enrolment data about users. web usage mining (wum) takes a gander at the aftereffects of customer relationships with a web worker, including weblogs, click streams, and informational index trades at a website or a social event of related areas. wum performs three guideline steps: preprocessing, design extraction, and results in examination it (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). giovanna use a lodap (log data preprocessor) tool to do preprocessing of web log data (castellano et al., 2007; nasraoui et al., 2000). to investigate web log information, we use lodap, a product device that cycles web access information to eliminate immaterial log passages, recognize gets made by clients, moksud alam mallik. an efficient fuzzy clustering algorithm for mining user session ...| 84 and gather client gets into client meetings. every client meeting contains access data (number of visits, season of visit, and so on) about the pages seen by a client; as a result, it depicts that client's navigational behavior. the term "user identification" refers to the process of identifying unique users from online log data. generally, the log document in extended common log design gives simply the pc's ip address and the client specialist. user registration-required websites will include additional user login information that can be utilized to identify users. each ip address will be treated as a user if the user login information is not available. after this, we have to recognize user sessions. here we will partition the web log data file into diverse parts known as user sessions. every session is considered a single visit to a website. identification of client meetings from the weblog record is a convoluted errand. this information can be used as a contribution to an assortment of information mining calculations it (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). for clustering user sessions, we employ the fuzzy c-means clustering technique. here we need to randomly select initial cluster centers. the similarity measure is done based on the page visit time using fuzzy intersection and union. even after preprocessing noise is still present in the web log data. olfa defined the similarity between user sessions where compute preprocessing and segmentation of web log data into sessions. preprocessing of web log data and cluster user sessions can achieve using the fuzzy clustering technique. this will affect the clustering result and similarity measures (olfa nasraoui et al., 2008). zahid explains an existing web usage mining framework. it uses the fuzzy settheoretic approach in preprocessing and in clustering. it improves mining results when compared with the crisp approach in preprocessing and clustering. because the fuzzy approach matches more with a real-world scenario. it is using the fuzzy c-means algorithm for clustering (zahid et al., 2011; ansari et al., 2011). using a fuzzy c-means clustering technique, castellano hopes to divide website users into different groups and generate session clusters. preprocessing should remove noise up to maximum because it will affect remaining operations like session identification and clustering the sessions. the fuzzy setbased approach can solve most of the challenges listed above. fcm needs an initial random selection of clusters. this work focuses on designing “an efficient fuzzy clustering algorithm for mining user session clusters from web access log data". it improves the quality of clusters discovered (castellano et al., 2006). 3. method: proposed system here a new efficient fuzzy clustering algorithm that can proficiently mine client session clusters from web access log information is proposed. the calculation manages the least of medians while choosing group focuses. the strategy lessens mean squared mistakes and takes out the impact of anomalies. 3.1. input data 85 | international journal of informatics information system and computer engineering 2(2) (2021) 80-93 the essential information sources utilized in web utilization mining are the worker log documents, which incorporate web server access logs and application server logs. the input server log data is downloaded from the site https://filewatch. net. filewatcher is a ftp search engine that monitors more than two billion files on more than 5,000 ftp servers. the downloaded file name is "pa. sanitized access. 20070109. gz". a sample server log file entry is given below (table 1). table 1. sample server log file entry 1168300919. 015 the time of the request 1781 the elapsed time for http request 17. 219. 121. 198 ip address of the client tcp_miss/200 http reply status code 1333 bytes send to the server in response to the request get the requested action http://www. quiethits. com/hitsurfer. php direct/204. 92. 87. 134 uri of the item being mentioned, customer client name, the hostname of the machine where we got the solicitation, text/html content-type of the object. 3.2. data mining every hour, well-known websites generate gigabytes of online log data. managing such massive records is a difficult task. log record sizes can be reduced by performing information cleansing, allowing mining assignments to be lifted. when a user requests for a web page enters or clicks on a url usually a single request will cause several urls to be generated like figures, scripts, etc. so all urls with a graphic extension should be removed. web robots are also identified and their queries are removed during data cleaning. in weblog data, a web robot (also called as web wanderers, crawlers, or spiders) generates numerous request lines automatically. robot’s request is unwanted because it is not generated by the user, it is generated by the machine. so, we should remove robot requests as removing them will increase the accuracy of clustering results. here we employed two methods for extracting robot requests. the first one is checking for an entry in "robots. txt" in http://www.quiethits.com/hitsurfer.php%20-%20%20direct/204.92.87.134 http://www.quiethits.com/hitsurfer.php%20-%20%20direct/204.92.87.134 http://www.quiethits.com/hitsurfer.php%20-%20%20direct/204.92.87.134 moksud alam mallik. an efficient fuzzy clustering algorithm for mining user session ...| 86 web log data and the second one is removing head requests (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). next is the removal of urls with query strings. normally url with query strings is used for requesting extra details from within the web page within the same session. since they are unnecessary, we will remove them as well (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). the input file is 30. 6mb in size and has 2,06,914 entries. after removing urls with graphic contents, the log file has 72,498 entries which are almost one third of the input file. after removing the web robot request, we have 72,305 entries. after removing urls with query string,we have 59,054 entries in the log file. then we will encrypt ip address to hide the user’s identity and to have ease in future processing and the ip address will be put away in a map with its encoded id. furthermore, each url will be appointed a unique number and it will be put away in a url map along with its number (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). the data cleaning algorithm is demonstrated in the following scheme: 1. step 1: remove each line of the input file one by one. 2. step 2: remove all urls with suffixes recorded in the above suffix list. 3. step 3: remove all urls produced by web robots. 4. step 4: remove urls with query strings. 5. step 5: take out the ip address and store it on a map. 6. step 6: code url with url number and store it on a map. 7. step 7: sort each line based on the ip address encryption code. 8. step 8: print in the required fields to a yield file. the output file after applying the above algorithm will be as shown in table 2. the output file is sorted in ascending order based on the encoded value of the ip address (table 2). table 2. output file after data cleaning ip time elapsed time bytes url ip1 1168300931. 828 142 1599 1 ip1 1168300935. 244 501 1617 2 ip1 1168300936. 604 1 1617 3 ip1 1168300941. 345 2 1593 4 ip1 1168300957. 585 186 1585 6 ip1 1168300985. 665 145 1563 10 87 | international journal of informatics information system and computer engineering 2(2) (2021) 80-93 3.3. user identification after cleaning input web log data, we can distinguish users. since the log file doesn’t contain user login information, we consider each ip as a user. next, we separate all solicitations identifying with the individual user. the algorithm for user identification is shown in the following scheme (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). step 1: split every line in the input file into obliged fields. step 2: store it(i. e. obliged fields) in a map m1 with ip address as the key and another map m2 as the worth. key of the map m2 is the time and worth is whatever is left of the fields. step 3: sort the internal map m2 considering the time key. step 4: print the content of the map m1 to the yield record. the organization of the yield document produced after user identification is shown in table 3. 3.4. session identification client session distinguishing proof is the technique of dividing the customer activity log of each customer into sessions, each addressing alone visit to the site. sites without client verification data generally depend on heuristic strategies for sessionization. the sessionization heuristic guides in isolating the genuine game plan of exercises performed by one customer in one visit to the site. keeping in mind the end goal to recognize client sessions we can try different things with two distinctive time-oriented heuristics (toh) as portrayed underneath (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011): toh1: the time term of a session should not surpass a limit α. let the timestamp of the main url demand, in a session be, t1. if another url asks for a session with timestamp ti it is allotted to the same session if and only if ti-t1≤ α. the principal url asking for with timestamp bigger than t1 +α is taken as the first request of the following session (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011): 1. step 1: the given steps ought to be finished for every line in the information input file. 2. step 2: if the line contains user id, then userid =user id of the line. 3. step 3: print line to output file under this user id and the first session of same user id. 4. step 4: in case that l is the first accessed log of the user then t1 = line. time else t2 = line. time. 5. step 5: if t2-t1≤ α at that point print line under the same session to the file. 6. step 6: if it is not as in the previous step i. e. step 5 then output user id and corresponding line under a new session, t1 = line. time. detailed information is shown in table 3. moksud alam mallik. an efficient fuzzy clustering algorithm for mining user session ...| 88 table 3. algorithm to create user sessions taking into account toh1. user time elapsed time bytes url ip1 1168300931. 828 142 1599 1 1168300935. 244 501 1617 2 1168300936. 604 1 1617 3 . . . . . . . . . ip2 1168300953. 645 648 260 5 1168300990. 665 143 260 14 toh2: the time spent on a page visit should not surpass a limit α. let a url that is most recently given to a session having a timestamp ti. the next url’s request fits in with the same session if and only if ti+1-ti≤ α where ti+1 is the timestamp of the new url’s request. this url is now the first of the following session (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). in our implementation for the interim, we are utilizing toh1. we have chosen 30 minutes as the estimation of the limit time. the algorithm for user session identification is shown in table 4 and the output file of session identification are shown in table 4. 3.5. dimensionality reduction removing to separate the logs references to low bolster urls (i. e. that are not bolstered by a predetermined number of user sessions) can give a powerful dimensionality decrease system while enhancing clustering. to implement this, we are removing urls that occur only once (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011) (see table 4). table 4: output file of session identification user session time elapsed time bytes url ip1s1 1168300931. 828 142 1599 1 1168300935. 244 501 1617 2 1168300936. 604 1 1617 3 ip1s2 1168302738. 407 81 1623 482 1168302745. 477 138 1559 483 . . . . . . . . . ip2s49 1168300953. 645 648 260 5 . . . . . . . . . 89 | international journal of informatics information system and computer engineering 2(2) (2021) 80-93 3.6. session weight assignment the session files can be divided for the clustering process in order to remove small sessions with the purpose of removing variation from the data. in any event, deleting these little measured sessions directly may result in the loss of a vital measure of information, especially if the number of these small sessions is significant. here we allot weights to every one of these sessions considering the number of urls got to by the sessions. session weight assignment is done based on the following equation (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). 𝑊𝑠𝑖 =0,if|𝑠𝑖|≤1 𝑊𝑠𝑖 =1, if|𝑠𝑖|≥1 where |𝑠𝑖| is the number of urls accessed in a particular session. 3.7. development of user session matrix here we represent sessions using a matrix. every row denotes a session, and the column denotes a url. if a url arrives in a session, then the entry for that url in the specific session will be more prominent than zero. it will be many events of that url in that session. if url is not present, then that entry will be zero. sessions are referred to by utilizing a sparse matrix in row-major form. it reduces processing time up to a great extent. after all, we are dividing to standardize the session matrix for every column by its greatest value (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011). for fuzzy clustering structures, the fuzzy c-means technique is commonly employed nowadays. so, in order to compare our new algorithm against the previous system, we used fcm (zahid et al., 2011; ansari et al., 2012; babuy et al., 2011; bezdek et al., 1984). 3.8. implementation of proposed system. the suggested system can be implemented as a fast fuzzy clustering technique for mining user session clusters from web log data, as described in section 5 titled "session clustering." the following is the primary part of the processing: at first, we will take one meeting say s1 and discover the distance between this meeting to every other meeting (say s2; s3; s4; …; sn)multiplied by the enrollment capacity of s1 to bunch focus 1(v1). next, we will sort these qualities into rising requests and take the middle. the above step will be done for all sessions s1; s2; s3; s4; …; sn. now these medians obtained from the above steps will be sorted and the least value will be taken. the session relating to the least worth will be taken as the main group community in this round. all above advances will be proceeded for bunch focus 2 up to group focus c(v1; v2; v3; …; vc). in this way, we will get new arrangements of bunch focuses in one round. new group communities will be determined up to a particular number of rounds till we get ideal bunch habitats. 3.9. modification in proposed system. moksud alam mallik. an efficient fuzzy clustering algorithm for mining user session ...| 90 here for every cluster center, we will be selecting the smallest value of medians. however, the issue is that abruptly we are getting the same smallest median in each iteration. so, in each cycle, we are getting the same cluster center repeatedly. so, we rolled out a little improvement in this algorithm. instead of selecting the least median in each round, we will choose the smallest median in the first round, the second smallest median in the second round, the third smallest median in the third round, and so on. by actualizing in this manner, we are demonstrating indicators of progress in the suggested algorithm's execution, which is superior to fcm. 3.9.1. fuzzy membership function expect to be x = {x1; x2; :::; xm} is the arrangement of information focuses or sessions. each point is a vector of the structure i = 1… m , xi = {xi1; xi2; :::; xin}. let v ={v1; v2; :::; vc} is a bunch of n dimensional vectors compares to c group habitats and each bunch place is a vector of the structure 8j = 1:::n , vj = {v1j; v2j; :::; vnj}. let uij addresses enrollment of information point(or meeting) xi in bunch j. the m×c enrollment framework u = [uij] shows portion of sessions to different bunch communities. it fulfills following models. ∑ 𝑢𝑖𝑗 = 1; ∀i = 1 … m 𝑐 𝑗=1 0<∑ 𝑢𝑖𝑗 𝑚 𝑖=1 <m, ∀j = 1 … c the participation esteem is determined by utilizing the accompanying equation (bezdek et al., 1984). uij= 1 𝑑 𝑖𝑗(xi,vj) 2 ( 1 𝑚−1 ) ∑ 1 𝑑 𝑖𝑗(𝑥𝑖,vk) 2 𝑐 𝑘=1 1 (𝑚−1) …………………(1) the initial cluster centers are chosen at random from the available sessions. then, using the equation for 𝑢𝑖𝑗 , the membership value of each cluster is calculated. the following equation can be used to calculate the euclidean distance between various data points and cluster centers (malik et al., 2021; bezdek et al., 1984). 𝑑 𝑖𝑗(𝑥𝑖,𝑣𝑗)=∑ 𝑑 𝑖𝑘(𝑥𝑘, 𝑖 𝑣 𝑘 𝑗 ) 2𝑛 𝑘=1 2 …………(2) where n is the number of dimensions of each data point, 𝑥𝑘, 𝑖 is the value of 𝑘𝑡ℎ dimension of 𝑥𝑖 , and 𝑣𝑘 𝑗 is the value of 𝑘𝑡ℎ dimension of 𝑣𝑗 which is the jth cluster center. 3.9.2. cluster center calculation the following formula calculates new cluster centers in each iteration step: di = median{(𝑑𝑖𝑗 (𝑠𝑘 𝑠𝑖) *𝑢𝑖𝑗 )}; ∀i ≠ k; k = 1…n p = argmin{(di:n); ∀i = 1…n} vj = sp at first, we will take one meeting say s1 and discover the distance between this meeting to every other session (say s2; s3; s4; …; sn) duplicated by the enrollment capacity of s1 to group focus 1(v1). next, we will sort these qualities in climbing requests and take the middle. the above advance will be accomplished for all meetings s1; s2; s3; s4; …; sn. presently from these medians got from above advances least worth will be taken. the 91 | international journal of informatics information system and computer engineering 2(2) (2021) 80-93 meeting relating to the least worth will be taken as the primary group place in this round . all above advances will be proceeded for bunch focus 2 up to group focus c(v1; v2; v3; ::::vc). in this manner, we will get new arrangements of the group focuses in one round. new group habitats will be determined up to a particular number of rounds till we get ideal bunch communities. 3.9.3. objective function calculation the target work is utilized to quantify the exhibition of the grouping calculation. the bunch habitats which are having less incentive for the target capacity will give smaller groups or better grouping results. the presentation record is determined utilizing the accompanying target work: oi=∑ (uij ∗ dij(vj − si))𝑐𝑗=1 ………. (3) 3.9.4. algorithm termination the means in these areas will have proceeded till a predetermined number of steps or till we get the base incentive for the goal work. the relating worth of the group places will be taken as last. 4. results and discussion for both fcm and the proposed fuzzy algorithm (fclm), we will give similar data information. that is, we will give a similar basic bunch habitat and ascertain the xie-beni index, partition coefficient, fs index, deviation, compactness, and separation of the group focuses overall. likewise, we will process the fluffy smallness and fluffy deviation of individual group habitats. after that, we will look at file esteems obtained by both fcm and the proposed calculation. the calculation which is getting less regard for legitimacy record performs preferred other over partition coefficient. for parcel coefficient, the calculation which gets a higher worth performs better (its most outrageous worth is "one"). detailed results are shown in figures 2-10. figure 2. effect of xie-beni index on value figure 3. effect of fs index on value figure 4. compactness moksud alam mallik. an efficient fuzzy clustering algorithm for mining user session ...| 92 figure 5. deviation figure 6. partition coefficient figure 7. fuzzy compactness of indvidual clusters figure 8. index value comparison for “for 450” figure 9. no of clusters figure 10. for 60 input cluster centre 19 5. conclusion the proposed fuzzy clustering algorithm further develops grouping. it decreases the impact of mean squared blunder and disposes of the impact of exceptions and consequently the clamor. so, we are improving bunching arrangements than in the fuzzy c-means algorithm. with better informational collections we can improve arrangements in our calculation. both fcm and proposed fuzzy clustering algorithm exhibitions are affected by starting group place decisions. to get away from this adverse consequence we can apply the mountain thickness capacity to pick basic group habitats. so we will get fitting beginning group habitats. in 93 | international journal of informatics information system and computer engineering 2(2) (2021) 80-93 preprocessing additionally, we need to do many enhancements like applying new techniques to channel robot demands. references ansari, z., azeem, m. f., babu, a. v., & ahmed, w. (2015). a fuzzy approach for feature evaluation and dimensionality reduction to improve the quality of web usage mining results. arxiv preprint arxiv:1509.00690. ansari, z., azeem, m. f., babu, a. v., & ahmed, w. (2015). a fuzzy clustering based approach for mining usage profiles from web log data. arxiv preprint arxiv:1509.00693. ansari, z., babuy, a. v., ahmed, w., & azeemz, m. f. (2011, september). a fuzzy set theoretic approach to discover user sessions from web navigational data. in 2011 ieee recent advances in intelligent computational systems (pp. 879-884). ieee. bezdek, j. c., ehrlich, r., & full, w. (1984). fcm: the fuzzy c-means clustering algorithm. computers & geosciences, 10(2-3), 191-203. castellano, g., fanelli, a. m., & torsello, m. a. (2006, september). mining usage profiles from access data using fuzzy clustering. in the 6th wseas international conference on simulation, modelling and optimization, portugal. castellano, g., fanelli, a. m., mencar, c., & torsello, m. a. (2007, november). similarity-based fuzzy clustering for user profiling. in 2007 ieee/wic/acm international conferences on web intelligence and intelligent agent technologyworkshops (pp. 75-78). ieee. castellano, g., mesto, f., minunno, m., & torsello, m. a. (2007, july). web user profiling using fuzzy clustering. in international workshop on fuzzy logic and applications (pp. 94-101). springer, berlin, heidelberg. cooley, r., mobasher, b., & srivastava, j. (1997, november). web mining: information and pattern discovery on the world wide web. in proceedings ninth ieee international conference on tools with artificial intelligence (pp. 558-567). ieee. han, j., kamber, m., & pei, j. (2012). data mining concepts and techniques 3rd edition vol. 3rd. mallik, m. a., zulkurnain, n. f., nizamuddin, m. k., & aboosalih, k. c. (2021, february). an efficient fuzzy c-least median clustering algorithm. in iop conference series: materials science and engineering (vol. 1070, no. 1, p. 012050). iop publishing. nasraoui, o., frigui, h., krishnapuram, r., & joshi, a. (2000). extracting web user profiles using relational competitive fuzzy clustering. international journal on artificial intelligence tools, 9(04), 509-526. ralph et al. development of pec based electrical load schedule software | 44 development of pec based electrical load schedule software ralph g. alvaro, eric kenneth b. cabrera, raineir ryan g. gutierrez, joshua d. punzalan, louie g. serrano, alma l. tangcuangco department of electrical engineering, don honorio ventura state university, philippines e-mail: raineirryangutierrez@gmail.com a b s t r a c t s a r t i c l e i n f o this paper presents the development of pec-based electrical load schedule software for residential electrical services design. all the sources that will be used for the computation and data will be based on pec. the testing of the project will be carried out in accordance with pec as well as other approved codes. calculating tools to help you understand the process of accurately sizing a commercial electrical panel based on proposed loads. to emphasize, accurate load sizing is best performed by an electrical engineer, which is reviewed by the building department in your jurisdiction. the software was developed using visual basic (vb. net) programming language. vb.net, or visual basic, is a development of the basic programming language. it is intended to be used in conjunction with an integrated development environment (ide). prior to the development of ides, programming languages like basic relied significantly on the dos command-line. a load schedule for different operating scenarios will show when peak consumption occurs and provide an opportunity to find out if all the high loads must operate at this time. it serves as a valuable calculation tool for electrical engineers, students, and technicians by providing a faster, easier, and more accurate means of carrying out some basic calculations, such as determination of the size of wire; conduit sizes; circuit breaker ratings; the main feeder size of wires; main circuit breaker ratings; voltage drop; and riser diagram using philippine electrical code tables. the results of these calculations help the designer to make vital decisions such as cable sizing and nominal ratings of protective devices required by each circuit and by the entire installation, in line with appropriate standards and regulations. article history: received 11 may 2022 revised 20 may 2022 accepted 25 may 2022 available online 26 june 2022 ___________________ keywords: diagram, electrical, singlephase, software, power system international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 45-58 ralph et al. development of pec based electrical load schedule software | 46 1. introduction the electrical load schedule computation is a summary of the data that helps electrical installers to know the equipment ratings and their contribution to the overall power consumption in an installation. the data in the electrical load schedules is usually backed by calculations which are based on standard electrical codes (pec) philippine electrical code. some of the electrical engineers compute their electrical load schedule manually and there are a portion who use microsoft excel and create a sheet for their computation. nowadays with the help of a fast phase advancement of technology many develop software with the means of improving their efficiency and accuracy. there are books to read where we can learn and understand the electrical load schedule, also known as pec or philippine electrical code, which is done by an electrical engineer. to have a precise and acceptable result, the purpose of this study is to create software that electrical engineers may use to identify everything they need to do and put everything in place. application of electrical products and equipment for additional installation or replacement is suggested to be consulted with a licensed electrical practitioner before installation for safety. failure to do so may result in fire, serious injury, or death. the electrical load schedule software will aid professional electrical engineer and electrical engineering in analysing load schedules. this program can compute the size of wire, conduit sizes, circuit breaker rating, the main feeder size of wires, main circuit breaker ratings, voltage drop and riser diagram using philippine electrical code tables. the software can be used for calculation and estimation for electrical engineers and master electricians in the project that has been professionally prepared and printed the electrical load schedule software combines a graphical user interface and the intelligence of software electrical load schedule for (bungalow/family home, offices, and administration building). to easily design, calculate and analyse low and medium voltage panel load schedules and distribution of panel boards. the electrical load schedule software is done by using a program that will calculate electrical loads to be used in an electrical calculation. this program will be using visual basic (vb.net), there are many programs language readily available in the market like c++, c#, java, and others. since the group’s knowledge in programming is limited, visual basic (vb.net) is another option (morrison & cornell, 2008). it is easier to use and you can develop something fairly fast. vb may also be used for web and application development, similar to java but with a different approach. the philippine electrical code (pec) is used nationally as the basis for safeguarding people, buildings, and contents from hazards that may arise from the use of electricity. this code contains provisions that are necessary for safety and thus is used as a basis for the legal enforcement in the installation of electrical system design and pec calculations are based on a 120-230 v ac source and an operating frequency of 60hz and the wiring sizes exist from each other’s table. this study aims to develop software for electrical engineers. the electrical load schedule software combines a graphical user interface and the data 47 | international journal of informatics information system and computer engineering 3(1) (2022) 45-58 intellect of pec to lighten, calculate, design, and analyse electrical load schedules and distributions of panel boards. the philippine electrical code (pec) is the basis of making sure the electrical service is sized properly and computed correctly. ● to provide these calculating tools as a guide to assist you to understand the process of accurately sizing a residential unit electrical panel based on proposed loads. ● to enlighten the results of the calculations are based upon the accuracy of the information that users provide. ● to emphasise the accurate load sizing is best performed by an electrical engineer which is reviewed by the building department in your jurisdiction. according to, the software was found to be reasonably effective and efficient in carrying out the required electrical design calculations at computer speed and accuracy (olatomiwa & alabi, 2012). the software serves as a valuable design tool for electrical engineers, students, and technicians by providing a faster, easier, and more accurate means of carrying out some basic calculations. the calculator software was created utilizing the java programming dialect and it serves as an important planning device for electrical engineers, students, and professionals by giving a speedier, less demanding, and more precise means of carrying out a few essential calculations, such as; assurance of the number of lighting installations required in a room; the plan current and control required per last sub-circuit; the whole associated load; voltage drop over chosen cables; and load adjusting over the three stages of a three-phase provided building. the results of these calculations offer assistance to the architect in forming crucial choices such as sorts of lights, sizes of cables, and nominal appraisals of defensive gadgets required by each circuit and by the whole establishment in line with fitting measures and directions. the software is called calculation for home electrical services. residential electrical services design calculator software. the software was created using the java programming language and serves as a useful design tool for electrical engineers, students, and technicians by making basic calculations such as determining the number of lighting fixtures required in a room, the design current and power required per final sub-circuit, the total connected load, voltage drop across chosen cables, and load balancing across the three axes faster, easier, and more accurate. the results of these calculations assist the designer in making critical decisions such as the types of luminaires, cable diameters, and notional protective device ratings required by each circuit and by the total system in accordance with acceptable standards. the few calculators software that provide residential design calculations utilise an approach that presume the home's load need is based on its area. this study, on the other hand, proposes a model that allows the designer to deal with precise loads predicted or needed by the customer while taking into account unusual load characteristics such as power factor. this program estimates apparent power, which is the amount of power required from the supply to the end user, while also accounting for variations in real load utilisation. the designer may enter all quantitative design elements utilised in the computations (diversity factor, utility factor, maintenance factor, and power ralph et al. development of pec based electrical load schedule software | 48 factor), improving design flexibility (parmar et al., 2012). calculating electrical loads needs to be improved to improve the accuracy of the data that affect the technical and economic indicators of the overall corporate complex. (khomiakov & ustinov, 2019). to reduce the inaccuracy in predicting electrical loads to acceptable levels in engineering calculations, the effect of energy-saving measures made on each electric receiver independently, particularly from the frequency converter installation, must be considered. the experiments undertaken confirm the necessity to alter the coefficients used in the calculations using the existing method. the analysis of the electrical load schedule takes far too long. there are steps to take, preparations to make, and requirements to meet. c.a. cole claims that while there are many various ways to construct an electrical design, there is only one item that cannot be changed (khomiakov & ustinov, 2019). the requirements of the code must be observed. according to polydorides, n., & lionheart, w. r. (2002), microsoft's vb.net is a programming language that runs on the.net framework. visual basic may be readily developed by any programmer that is familiar with vb.net. the common language runtime environment for microsoft's visual studio program is used by vb.net. the cost of a schedule is determined by the total energy consumption of the equipment, which is calculated on a regular basis. every 15 minutes, the power provider and the plant owner agree a target usage. any deviation from the contractual usage is punished, and the goal is to keep the total of these fines as low as possible. hait, a., & artigues, c. (2011) according to khuntia, s. r., rueda, j. l., & van der meijden, m. a. (2018), for utilities and planners, long-term electrical load forecasting is critical. the use of a multiplicative error model (mem) to anticipate long-term demand is proposed in this work. the advantages of taking volatility into account is demonstrated by out-of-sample forecast outcomes and directional accuracy during the great recession. for the distribution system planning problem, a novel technique based on the seeker optimization algorithm (soa) is given. soa entails the simultaneous installation of automatic reclosers (ras) to prevent brief failures. it may be utilized effectively to optimize the design of an advanced power distribution system. kumar, d., & samantaray, s. r. (2014) according to karabulut, k., alkan, a., & yilmaz, a. s. (2008), load management is a capacity that load distribution centers and electric companies demand. longterm load forecasting provides direction for installation maintenance and construction planning. power companies have been more interested in new forecasting approaches for more trustworthy and accurate estimates. short-term, mid-term, and long-term load forecasting assessments are available. what vb developers can achieve with visual basic.net is limitless, and books for visual basic.net must be created from the bottom up. vb developers are continuously seeking for publications that are clear and effective in 49 | international journal of informatics information system and computer engineering 3(1) (2022) 45-58 communicating these new issues. barnaby, t. (2002) 1.1 power load calculator this software will help you calculate the load on a circuit to see if it is excessive. you may also determine the size of circuit breaker required for the load. this is especially useful during pre-production since it allows you to quickly assess whether or not you'll require an external generator and how many you'll require. calculating the minimum required load for your power supply is a straightforward task. by glancing up a few ratings from the datasheet, you can immediately apply the proper load size (power load calculator, 2007-2022). 1.2 design master electrical it enables the user to design quality schedules with different electrical characteristics that are simple to manage and read. it enhances electrical calculations with new features. this application's design is simple and straightforward. it responds quickly when doing its role. the main part of designing electrical services is the userfriendly interface. design master electrical is a full electrical drawing and model solution that works directly in autocad, requiring no other software (design master software, 2001-2022). 1.3 wiresizecalculator.net the wire size calculator makes estimating the size of wires and cables for pump applications very simple and quick. this site has a number of complex calculators to help you size wire securely and accurately. it has calculators, tables, and tools to help you size wire and conduit correctly (wiresizecalculator.net, 2021). 1.4 necdesign.com nec calculator is a set of calculators that will correctly, efficiently, and consistently handle all of your basic electrical design jobs. wire size, conduit fill, fault current, power factor, transformer, underground cover, lighting design, motors, voltage drop, and more are all done in accordance with current national electrical code regulations. the reports are thorough, succinct, and wellwritten (nec design software, 2014). 2. method the following procedures were employed in the development of p.e.c. based electrical load schedule software: methods of program development the research methodologies used for the study are described with specific objectives to achieve the design conditions of the project. the projects start with the definition and function of the components based on the philippine electrical code (pec) (pec, 2017). all the sources that will be used for the computation and data will be based on philippine electrical code. then, testing of the project will be done (fig. 1). program language the vb.net or visual basic is a development of the basic programming language that is intended to be used in conjunction with an integrated development environment (ide). programming languages like basic relied significantly on the dos command-line prior to the development of ides. programmers had to write their code by hand and use dos to convert it into executable files, which was a timeconsuming procedure with a lot of chance for error. an ide, on the other hand, streamlines the process by offering a single interface that makes coding easier ralph et al. development of pec based electrical load schedule software | 50 and allows for one-click compilation into an executable file. we think visual basic is a very approachable programming language for beginners who want to write basic applications because of its simplicity. visual basic is a microsoftdeveloped programming language that is primarily used to create method programs in the windows system. electrical design based on philippine electrical code (pec) all the data based on pec will be used in the mathematical modelling of the software. this includes all the necessary computation in the load schedule, voltage drop and the riser diagram of the software such as ground wire of the main circuit breaker (table 2.50.3.17), size of wire (table 3.10.2.6), ground wire for the size of the wire (table 2.50.6.13), horsepower (air conditioning unit) (table 4.30.14.2), ampere ratings for the circuit breakers (table. 2.40.1.6) and ampere rating (conduit). all of this will be included in the mathematical modelling of the software. 2.1. methods of program development fig. 1. program development 2.1.1. program application this program application you will choose or select of what type of load schedule you will create. there are two option calculations, the single phase and the three phase. 2.1.2. input data the user will select the type of loads and input all the required data in creating a load schedule table. the data entry operator may be required to verify or edit data as it is entered. there will be a dropdown option when you insert your quantities. 2.1.3. construction of load schedule table loads that are inputted by the user are now then put in the load schedule. it includes a circuit number, types of loads, rating rated current, 51 | international journal of informatics information system and computer engineering 3(1) (2022) 45-58 circuit breaker ratings, sizes of wires, conduit sizes, voltage drop and riser diagram this will help designers to properly size the power backup generators or other alternative power sources. 2.1.4. perform computation the program performs two different types of computation: load schedule and riser diagram. for the load schedule table, the program computes the rated current and gives the sizes of wire, conduit size, voltage drop, circuit breaker ratings for each load. the load schedule table, computes the overall current and gives the main feeder sizes of the wires, conduit size, and circuit breaker rating. for the design of the riser diagram, the program only computes the circuit number of the breaker and the size of the wire grounding electrode. 2.1.5. voltage drop it will automatically show you the voltage drop. if volts drop, amps drop too. that is why when you find a voltage drop in a connection or cable, you know the connection or cable is restricted. 2.1.6. display results the program displays the computed results in the load schedule table and the riser diagram in its dedicated platform. the load schedule table shows the sizes of wires, conduit size, and circuit breaker rating of each load and also the main feeder sizes of wires and main circuit breaker ratings and voltage drop. while the riser diagram only displayed the circuit breaker number and grounding electrode. the process will handle the recommended inputs and it will be shown downwards while you put the necessary details in your computation such as the proper wire size and number of circuits to be used, types of loads, rating rated current, circuit breaker ratings, size of wires and conduits to be used in the project. 2.1.7. save if the user is finished with the design, the user can choose to save it or discard it. if the user chose to save the file, the program produces a text file for the riser diagram with load schedule computation. the software can provide printed output when you finish the design. 2.2. software formulas in the single-phase system this is the simplest. the kva may be calculated easily using the kw and power factor. the current is calculated by dividing the kva by the voltage. while in the three phase system the voltage is the major difference between a three phase and a single-phase system. in a three-phase system we have the line-to-line voltage (vll) and the phase voltage (vln). 2.2.1. conductor sizing branch circuit (lighting and appliance load) qty. x 100 (if lighting) = watts (as per pec standard) qty. x 180 (if con. outlet) = watts (as per pec standard) (section 2.20.2.5 pg. 55) tv, refrigerator, water pump and range etc. as per owner choice of wattage (va) va/v = a (amperes) ralph et al. development of pec based electrical load schedule software | 52 watts /volts = amp. load 2.2.2. motor load the current on the branch circuit feeding a single motor is taken as 125% the full load rated current of the motor. ic = ifl x 1.25 the current on the branch circuit feeding two or more is taken as not less than 125% of the full load current rating of the highest rated motor in the group plus the sum of the full load current rating of the remainder of the motors in the group. (article 2.10, pec 2017. pg. 43) 2.2.3. overcurrent protection rating branch circuit (lighting and appliance loads) the rating of the overcurrent protection should be capable of carrying the computed line current. it should be the nearest higher commercially available overcurrent protection/circuit breaker. the ampere trip of the overcurrent protection should not be higher than the ampacity of the conductor that is being protected. motor load the overcurrent device protecting the individual motor branch circuit must be able to carry the motor starting current. it must have a time delay to allow the motor to start and come up to speed. for single phase (all types no code letter) icb = 250% ifl – inverse time breaker icb = 700% ifl – instantaneous trip breaker iss = 300% ifl – non time delay fuse iss = 175% ifl – dual element for other types of motor with code letters. (article 4.30, pec 2017) feeder circuit (lighting and appliance loads) single or three phase loads the rating of overcurrent protection shall be capable of carrying the computed line current. it should be the nearest higher commercially available overcurrent protection should not be higher than the capacity of the feeder conductor protected. minimum ampacity of the conductor required (ic), full load current of the motor (ifl).= full load current of the highest rated motor (iflhrm). ic = 1.25% ifl hrm + ifl (article 2.10, pec 2017, pg. 43). single phase system service entrance conductors (size of wire) the service entrance is where the wires connected to the load side of the meter enter the house or building. the service entrance in a home is typically called to as a breaker or power supply. if = (ifl x d.f) + 25% (hml) service equipment (size of breaker) this required equipment, which typically consists of circuit breakers, switches, and fuses, as well as their equipment, is connected to the load end of service conductors that serve a structure and serves as the main control and cut off of the supply. imcb = (ifl x d.f) + 150% (hml) (size of breaker) (section 4.30.2.4 pec 2017, pg. 349,350) three phase system 53 | international journal of informatics information system and computer engineering 3(1) (2022) 45-58 delta connected system service entrance (size of wire) 𝐼𝐹 = [∈ 𝐼3∅ + √3 ∈ 𝐼1∅𝐻 + .25 𝐼ℎ𝑟𝑚][𝐷𝑒𝑚𝑎𝑛𝑑 𝐹𝑎𝑐𝑡𝑜𝑟] service equipment (size of breaker) 𝐼𝑐𝑏 = [∈ 𝐼3∅ + √3 ∈ 𝐼1∅𝐻 + 1 .5 𝐼ℎ𝑟𝑚][𝐷𝑒𝑚𝑎𝑛𝑑 𝐹𝑎𝑐𝑡𝑜𝑟] a delta connection is a three-phase system connection in which three elements in series form a triangle, with the supply being input and output at the three junctions. minimum ampacity of the feeder conductor (if), summation of line currents (il), current of the highest rated motor in the group/panel (ihrm), summation of 3phase line current (∈i3∅), summation of the single-phase line current (∈i3∅h). voltage drop is calculated using the most universal of all electrical laws: ohm's law. this states that the voltage potential across the conductor is equal to the current flowing through the conductor multiplied by the total resistance of the conductor. in other words, vd = i x r. a simple formula was derived from ohm's law to calculate the voltage drop across a conductor. this formula can help you determine voltage drop across a circuit, as well as the size wire gauge you will need for your circuit based on the maximum desired voltage drop. the philippine electrical code (article. 2.15, 2017) states that the voltage drop of a feeder circuit must not exceed 5%, and the voltage drop of a branch circuit must not exceed 3%. the voltage drop formula illustrates whether the power given by the voltage source is condensed as electric current runs through the parts of the electrical circuit that do not supply the voltage. furthermore, voltage drops between the source's internal resistances and connections are undesirable since supply energy is wasted. furthermore, voltage drop across active circuit parts and loads is preferable since provided power performs a competent task. voltage drop formula general formula vd = iload zwire where: 𝑍 = √𝑅2 + 𝑋2 singlephase circuits 𝑉𝐷 = 2𝐼𝑍 ( 𝐿 305 ) ( 1 𝑁 ) threephase circuits 𝑉𝐷 = 𝐼 √3 𝑍 ( 𝐿 305 ) ( 1 𝑁 ) where: z is the wire impedance, ω l is the total length of wire, m n is the number of sets (section 10.1.1.9 alternating-current resistance and reactance pg. 942). conductors for feeders, as sized to prevent a voltage drop exceeding 3 percent at the farthest outlet of power, heating, and lighting loads, or combination of such loads, and where the maximum total voltage drop on both feeders and branch circuits to the farthest outlet does not exceed 5 percent, will provide reasonable efficiency of operation. 2.15.1.2 minimum rating and ralph et al. development of pec based electrical load schedule software | 54 size (feeders), 2.10.2.2 minimum ampacity and size (branch circuit) pec. 2.3. research tool evaluation form. an instrument is used to assess the operational feasibility of the system. the following criteria were provided in order to evaluate the developed application. during the evaluation of the developed system, the researchers distributed questionnaires and feedback forms through an online form to the respondents of the system and they were evaluated by a licensed electrical practitioner. the developed application was evaluated by the target users. in addition, computed means are interpreted in the five-point likert scale, which uses a rating scale ranging from not at all satisfied (1) to very satisfied (5) and the conversion of the qualified weighted mean into a descriptive rating. the researchers used the product quality model defined in iso/iec 25010 for the study. the researcher went through eight primary phases in order to execute this model: functional suitability, performance efficiency, compatibility, usability, reliability, security, maintainability and portability. functional suitability is when used in certain conditions, this characteristic demonstrates the extent to which a product or system provides functions that meet stated and implied expectations of the electrical load schedule software while also adhering to the philippine electrical code's standards. later, the requirements of the electrical load schedule software and the pec standards guided the creation of the template and layouts. this feature indicates performance in relation to the quantity of resources consumed under specified conditions. then the compatibility, the extent to which a product, system, or component may share information with other products, systems, or components while sharing the same software environment. for usability, the degrees to which a product or system may be utilised by specific users to accomplish specific goals with effectiveness, efficiency, and satisfaction in a specific context of usage. then reliability, the degree to which a system, product, or component fulfils defined functions under given conditions during a specific time period. security is the degree to which a product or system safeguards information and data so that people or other goods or systems have the proper level of data access for their kinds and degrees of permission. the degree of efficacy and efficiency with which a product or system may be adjusted to enhance, rectify, or adapt to changes in the environment and needs is represented by the maintainability characteristic. finally, portability describes the ability with which a system, product, or component may be transferred from one software or other operational or consumption environment to another. 3. results and discussion this chapter presents the major findings of the study. 3.1. software homepage fig. 2. shows the electrical load scheduling program. the software's logo will be shown on the welcome page. as for the image shown you can also see the username and password wherein the user will put its own username and password then click the login button to continue if not the user can just click the 55 | international journal of informatics information system and computer engineering 3(1) (2022) 45-58 exit button, then the software will continue to exit. fig. 2. welcome page of the program 3.2. calculation option page the single-phase and three-phase calculations are the two possibilities for computation on this menu. the table will now be displayed, where you may make a list of all the scheduled loads in the installation. get the electrical characteristics for each load in a single phase, such as wire size, conduit size, circuit breaker ratings, main feeder wire size, main circuit breaker ratings, voltage drop, and riser diagram. in all three phases, the procedure is the same, but the computations are different. when everything is finished, the riser diagram will be shown (fig. 3). fig. 3. calculation option page 3.3. load description an outlet is a place on the wire system that supplies power to electrical equipment. a lighting outlet is one designed for electrical use. on household general-purpose branch circuits, there is no limit to the number of light or receptacle outlets. on the software, the following lights will be visible; halogen lamp, led dual floodlight, fluorescent lamp, incandescent lamp, mercury lamp, led emergency light, hps lamp, led lamp, spot light, led linear lamp, compact fluorescent, led high bay lamp, incandescent lamp, led outdoor light, pin light, and smart bulbs. this is due to their high-power consumption or other fundamental properties such as high initial current and frequency of use. appliances are typically seen as a single piece of equipment used to perform household tasks in the home. the following appliances will be shown in our program; exhaust fan, hand dryer, freezer, fan coil unit, ceiling fan, television, heater machine, led tv, water heater, washing machine, water pump, air conditioner unit, dryer, and refrigerator. ralph et al. development of pec based electrical load schedule software | 56 a convenience outlet is a type of outlet that is mounted on a room's wall and is used to power electrical appliances. the electricity from the incoming source is connected to the line side of an outlet. power exits the device and continues down the load side of the circuit. the predicted load for a single 230-volt outlet. here’s the following; convenience outlet, range, duplex convenience outlet, spare, and spo (fig. 4). fig. 4. load description 3.4 respondents of the evaluation table 1 presents the respondents of the study. the following are the results taken from the evaluation of 7 registered electrical engineers (ree) and 1 professional electrical engineer. product quality defined in iso/iec 25010 with results of the evaluators. table 1. respondents of the study respondents frequency percentage professional electrical engineer (pee) 1 12.5% registered electrical engineer (ree) 7 87.5% total 8 100% table 2 shows the overall result of the iso/iec 25010. from the data that the researchers collected and after analysing and reviewing the respondents’ answers, it shows that from 8 respondents who participated in answering the survey questions, 35.89% answered more than satisfied to use the pec based electrical load schedule software for their design and to be further enhanced. moreover, 34.68% are satisfied and agree to use the software. meanwhile, 18.55% answered that they are fine with whether the system will be built or not. furthermore, 10.48% answered that they were very satisfied and wanted the software to be further enhanced to give more pleasing 57 | international journal of informatics information system and computer engineering 3(1) (2022) 45-58 satisfaction to the user. and last 0.40% who are not at all satisfied. table 2. overall result of the iso/iec 25010 indicators respondents m e a n m e d ia n m o d e standar d deviatio n response descriptio n 1 2 3 4 5 functional suitability 0 5 10 8 1 3.21 3 3 0.83 satisfied performance efficiency 0 6 10 5 3 3.21 3 3 0.98 satisfied compatibilit y 0 3 6 7 0 3.25 3 4 0.77 satisfied usability 0 10 13 19 6 3.44 4 4 0.97 satisfied reliability 0 6 9 11 6 3.53 4 4 1.02 more than satisfied security 0 7 14 16 3 3.36 3 4 0.87 satisfied maintainabi lity 1 7 14 14 4 3.33 3 3.4 0.97 satisfied portability 0 2 10 9 3 3.38 3 3 0.82 satisfied total 1 46 86 89 26 3.32 3 3 0.93 satisfied 1 2 3 4 5 indicato rs respondents m e a n m e d ia n m o d e standar d deviatio n response descriptio n 1 2 3 4 5 1 46 86 89 26 3.38 3 4 0.92 satisfied 4. conclusion and recommendations conclusion the program was found to be relatively effective and efficient when it came to doing the needed electrical design calculations at computer speed and accuracy. the program is a helpful design tool for electrical engineers since it allows them to perform certain ralph et al. development of pec based electrical load schedule software | 58 fundamental calculations faster, easier, and more accurately for home electrical design work. recommendations the researchers should refer the program to autocad. the computation of the electrical load schedule, as illustration, is direct to convert to autocad. we'd want to use our software to build it, but we wouldn't have enough time. since the group was mainly focused on bungalows or family homes, offices, and administrative buildings, this subject suggests that it contains more industrial structures where suitable demand variables for various industrial buildings must be addressed. references [1] morrison, j., & cornell, g. (2008). programming vb. net: a guide for experienced programmers. a press [2] olatomiwa, l., & alabi, a. c. (2012). design and development of calculator software for residential electrical services design. [3] khomiakov, k. a., & ustinov, d. a. (2019, october). improving the accuracy of calculations of electrical loads for industrial enterprises. in journal of physics: conference series (vol. 1333, no. 6, p. 062009). iop publishing [4] parmar, r., mangrola, m. h., parmar, b. h., & joshi, v. g. (2012). a software to calculate crystalline size by debey-scherrer formula using vb .net. multidisciplinary edu global quest, 1(1). [5] polydorides, n., & lionheart, w. r. (2002). a matlab toolkit for three-dimensional electrical impedance tomography: a contribution to the electrical impedance and diffuse optical reconstruction software project. measurement science and technology, 13(12), 1871. [6] hait, a., & artigues, c. (2011). on electrical load tracking scheduling for a steel plant. computers & chemical engineering, 35(12), 3044-3047. [7] khuntia, s. r., rueda, j. l., & van der meijden, m. a. (2018). long-term electricity load forecasting considering volatility using multiplicative error model. energies, 11(12), 3308. [8] kumar, d., & samantaray, s. r. (2014). design of an advanced electric power distribution systems using seeker optimization algorithm. international journal of electrical power & energy systems, 63, 196-217. [9] karabulut, k., alkan, a., & yilmaz, a. s. (2008). long term energy consumption forecasting using genetic programming. mathematical and computational applications, 13(2), 71-80. [10] barnaby, t. (2002). distributed. net programming in vb. net. apress. 61 | international journal of informatics information system and computer engineering 4(1) (2023) 61-74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 pharmacovigilance strategy for covid-19 vaccine information management and utilization among health information managers in kwara state, nigeria abdulsalam abiodun salman1, priscilla ayooluwa jacob1*, saleh idris2, abdullateef bako3 1department of library and information science, university of ilorin, nigeria 2senior librarian, abdullahi fodiyo library, usmanu danfodiyo university, nigeria 3department of urban and regional planning, university of ilorin, nigeria *corresponding email: priscillakolawole@gmail.com a b s t r a c t s a r t i c l e i n f o diversified opinions on the efficacy of covid-19 vaccine have necessitated investigation on the pharmacovigilance strategies used to manage covid-19 vaccine information and perception of health information managers. narrative research design was adopted while structured interview was used to collect data from respondents and qualitative data analysis technique was employed. total enumeration was adopted to gather data from eight purposively selected health information managers in the ministry of health, kwara state. the study revealed that majority of the respondents are aware of pharmacovigilance and covid-19 vaccine information are managed through who guideline. virtual meetings, emergency preparedness and response are some opportunities provided for covid-19 vaccines information management. vaccine hesitancy, non-compliance to usage-regulations, misinformation and overwhelmed facility were identified as challenges with acceptance of the vaccine. the study thus recommended that pharmacovigilance be enhanced among health information managers using advance information management technology. article history: submitted/received 18 jan 2023 first revised 12 feb 2023 accepted 18 mar 2023 first available online 14 apr 2023 publication date 01 jun 2023 aug 2018 __________________ keywords: covid-19 vaccine, health managers, information utilization, management, pharmacovigilance. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(1) (2023) 61-74 https://doi.org/10.34010/injiiscom.v4i1.9572 mailto:priscillakolawole@gmail.com salman et al., pharmacovigilance strategy for covid-19 vaccine information…| 62 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 1. introduction the recent outbreak of the globally dreaded disease, corona virus, which the world health organization (who) declared on the 11th of march 2020 as a pandemic has to a great extent threatened the global public health (cucinotta & vanelli, 2020; mahase, 2020). the corona virus also called covid-19 has infected and a led to the death of so many people across the globe. in response to the high rate of death, a number of measures were recommended, especially by who, to repress and prevent the spread of the disease. the government of countries all over the globe, nigeria inclusive, has thus swift into rapid action on prevention measures against the virus. the responses of various governments to the covid-19 pandemic have been multifaceted, depending on the intensity at which the virus hits individual countries; these include: outbreak management such as lock down policy (that is, suppression versus mitigation), building and sustaining flexible capacities to respond to viral outbreaks, ensuring sufficient clinical treatment resources for severe cases and implementing strategies to address the economic and psychosocial repercussions of the pandemic (wolf et al., 2020; zandvoort, 2020; onuh, 2021). however, the news of this virulent disease has activated a massive dissemination of health related information, most of which are about the infection, causes, symptoms, managing, adapting to the ‘new normal’ (corpuz, 2021) and ways to overcome when infected. though the information is at times sensational and distorted, it is being published through all media including print, non-print and especially social media. unfortunately, most of the information that goes viral from unprofessional sources is in opposition to those carried by health information managers and experts. this results in difficulty in citizens’ decision on use of recently approved covid-19 drugs and vaccines, imported into countries such as nigeria. the significance of vaccine is to assist in the fight against particular disease(s) by boosting the immune system of humans. therefore, information management strategy about newly deployed vaccine(s) becomes imperative, so as to ensure the right flow of information. likewise, it would help to investigated and scientifically managed people’s perception on use of the vaccine. findings of some studies have reflected mixed feelings on the efficacy and safety of some currently utilized drugs for the treatment of covid-19. some anecdotal evidences have shown that some drugs (such as, chloroquine, hydroxychloroquine, favipiravir, and umifenovir) approved for the treatment covid-19 in a country like nigeria (chen et al., 2021; heustess et al., 2021) constitute components that could lead to the ineffectiveness of covid-19 vaccine. effects of these rumors on people’s perception may include therapeutic errors. thus, influencing choice and decisions of individual patients on medication and vaccine use, particularly, those who are incessant users of the social media. therapeutic errors are changes in human behaviour caused by inappropriate information management strategy (aronson, 2009). these human behaviours may require that pharmacovigilance stakeholders activate risk minimization strategies particularly in relation to recently developed covid19 vaccines. https://doi.org/10.34010/injiiscom.v4i1.9572 63 | international journal of informatics information system and computer engineering 4(1) (2023) 61-74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 pharmacovigilance is principally concerned with drug safety, evaluation, collection, identification, monitoring, and prevention of ominous impacts of pharmaceutical products (al-mustansir, 2013; akarowhe, 2020). santosh and tragulpiankit (2011) further described pharmacovigilance as an ongoing process to monitor drug safety and make available new information about adverse drugs reactions. the concept of pharmacovigilance is on the basis of the advice by the strategic advisory group of experts (sage) in the oxford university, united kingdom on immunization, in order to save lives due to the outbreak of the covid-19 pandemic that has infected and terminated millions of lives across the globe (keni, 2020). as a matter of urgency, the need for information managers to examine the results of communication deficiency on people’s behavior, effects of the dangers arising from this conduct has become vital. also, it is necessary to investigate the rationale behind the claims of using vaccines vis-àvis the involvement of stakeholders in emergency management and strategy in order to deal with the covid-19 vaccine crisis particularly in kwara state, nigeria. 1.1. literature review pharmacovigilance encompasses the scientific and operational efforts involved in identifying, evaluating, comprehending, and mitigating the negative effects of drugs or other issues related to drug usage (akarowhe, 2020; tyagi, 2021). berlin (2005) declaration on pharmacovigilance noted the importance of sharing pharmacovigilance data for better reporting and information gathering. they also declared that pharmacovigilance is important for providing better information, collection of feedback (particularly from drug users and patients) so that governmental or non-governmental institution can commence appropriate studies on specific drug safety concerns). scott (2021) clarified that health information management involves the gathering, examination, safekeeping, and preservation of high-quality patient health information, which can be in the form of physical documents, a mix of physical and digital records (hybrid), or fully electronic health records. information management is very essential in healthcare delivery and particularly under the conditions of global pandemic. according to gilyarevsky (2019), affirm that the efficiency of information management (im) in critical situations can be determined by the capability of management systems to develop and implement a proactive strategy within a short time. a survey carried out by aleinikov, kurochkin and maltseva (2021) showed improvement in the communicative efficiency of theanti-crisis information management strategies in the context of the global pandemic. it was discovered that the anti-crisis information management strategies used in russia have accelerated digitalization of health information to help managers react to the spread of covid-19. thus, im has aid the process of pharmacovigilance in the country. however, the crises presented by these economic and life-threatening diseases such as the covid-19 pandemic to public and healthcare organizations are constantly escalating. in response, taiwan presented a set of operational guidelines for healthcare organizations to launch effective countermeasures against such crises (wang & wu, 2020). this was https://doi.org/10.34010/injiiscom.v4i1.9572 salman et al., pharmacovigilance strategy for covid-19 vaccine information…| 64 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 accomplished through the implementation of efficient knowledge management techniques. additionally, the utilization of information technology (it) applications greatly enhances organizations' crisis management and knowledge management capabilities by improving their responsiveness and adaptability. wang and wu (2021) emphasized the potential of innovative it-enabled tools, such as non-contact monitoring devices, intelligent robots, and telemedicine, in minimizing the risk of exposure and facilitating informed decision-making during the covid-19 crisis. the study findings ascertained the effectiveness it-enabled knowledge management practices in preventing and minimizing the undesirable consequences of the crisis in taiwan. in the same vein, the nigeria center for disease control (ncdc) with the help of health management institute, have set up guidelines to guide against the spread of infectious, communicable and killer diseases including ebola, lassa fever and covid-19. the ncdc which was established in nigeria in 2011 has been authorized to coordinate the surveillance of these communicable diseases (ncdc act, 2018; maxmen, 2019; erameh & ojakorotu, 2021). ncdc develop guidelines, protocols, surveillance systems, trains personnel at various levels, and shares public health information in order to provide leadership for the closely monitor infectious diseases in nigeria. however, it appears difficult for citizens to trust the information provided by ncdc (abdullateef & okonkwo, 2021). also, researchers have been interrogating the noticeable commercialization means by which nigeria has been responding to the ovid-19 pandemic. these include falsification of figures, reliance on foreign vaccines, replication of health care projects, scarcity of test, treatment kits and drugs, conflicting government reports and violations of rules by government officials and the perceived non-activity of the covid-19 taskforce (erameh & ojakorotu, 2021). furthermore, abdullateef and okonkwo (2021) recently investigated nigerian’s hesitation to trust ncdc’s information on covid-19 management using the survey research approach. the research findings indicated that misperceptions and skepticism towards the ncdc were present among all segments of the population, including those who are educated and have higher incomes. the main reasons for this lack of trust were identified as inconsistent and politicized reporting, lack of transparency regarding patients' data, and perceived tendencies towards corruption. similar misperception occurred in the united states of america (usa) when citizens perceived that covid-19 information management was politicised. for example, the us president stated that he would not wear a mask on the day the center for disease control (cdc) recommend that the public wear masks (yamey & gonsalves, 2020). the us cdc nearly became paralyzed as the administration tended to politicize its messaging while minimizing the agency’s scientific credibility. the study recommended increased openness, as the populace would support honest information, and access to persuading information about exercises to diminish the infection, clarity in treatment of covid-19 and budget spending. the national academy of sciences and national academy of medicine (2021) also expressed their concern over https://doi.org/10.34010/injiiscom.v4i1.9572 65 | international journal of informatics information system and computer engineering 4(1) (2023) 61-74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 ongoing reports and incidents on politicization of science, such as overriding evidences and advice from public health officials. the reports indicated that the public have been exposed to the danger of the pandemic and often misled through their inability to recognize the significance of the crisis and failure to abide with advice from experts. hussain et al. (2021), likewise observed a noticeable increase in positive sentiment of the public towards covid19 vaccines in the united states. this sentiment led to several discussion groups on the acceptability of the vaccine. as republican legislators became more engaged in public discussion on vaccine progress, there are increasing implications for covid-19 vaccine uptake (guntuku et al., 2021). for example, anti-vaccination allies have been deeply engaged in discussions and dissemination of misinformation and conspiracy theories (germani & billerandorno, 2021). kresps et al. (2021) in a survey on attitude of 1096 public members in america toward covid 19 vaccination examined the relationships between vaccine attributes, financial incentives, and misinformation on public immunisation preferences. the study revealed that higher degrees of vaccine efficacy significantly increased individuals’ willingness to receive the covid-19 vaccine. on the other hand, a high incidence of minor side effects of the vaccine, and emergency use authorisation decreased individuals’ willingness to take the vaccine; and vaccine manufacturers do not seem to influence public willingness to be vaccinated. from the study, there was evidence that belief in misinformation about covid-19 treatments was positively associated with vaccine hesitancy. these findings therefore have implications for public health strategies intending to increase levels of community vaccination. williams and dienes (2021) explored the public attitudes to covid-19 vaccines in uk through a qualitative study. the review identified three primary reasons for participants’ decision to be vaccinated against covid-19. delay due to a perceived need for more information and until vaccine was “required” in the future were revealed as reasons for vaccine hesitancy. the study also identified the perception of the public as vaccination being a social norm, necessity and trust in science as the facilitators of vaccination decision. furthermore, preference for natural immunity, concerns over possible side effects, distrust in government, perceived lack of information and conspiracy theories were some of the barriers to vaccine uptake. regardless of the facilitators and barriers posed to the vaccine uptake, globally, being vaccinated seems not to change people’s decision to adhere to covid-19 measures as pharmacovigilance entails. developing countries are not also exempted from these perceptions on covid-19 vaccine uptake. the existing situation in nigeria, for instance, suggests the need for strong collaboration among information managers and pharmacovigilance officials. this limitation hampers the national pharmacovigilance center's efforts to ensure drug safety and has led to inadequate awareness programs for pharmaceutical practitioners and other healthcare providers. vallano et al. (2005) suggest that while most doctors are aware of pharmacovigilance programs, there are still some who lack awareness. in addition to medical practitioners, some patients lack sufficient knowledge on https://doi.org/10.34010/injiiscom.v4i1.9572 salman et al., pharmacovigilance strategy for covid-19 vaccine information…| 66 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 how to ensure the safety of drugs when managing their health conditions. enger et al. (2013) suggest that patients often have limited understanding of patientoriented prescription drug information, which may be due to inadequate awareness through traditional information channels such as radio, television, internet, or social media. therefore, pharmacovigilance systems should strive to provide patients and practitioners with practical and actionable information about the use of medications, including drugs and vaccines (dal pan, 2014). information management systems are therefore needed for provision of valuable and reliable information, particularly in times of health crises. 1.2. objectives of the study the broad objective of this study is to determine the pharmacovigilance strategy used in determining of covid19 vaccine information management and utilization among health information managers in kwara state. specifically, the study sought to: i. investigates the level of pharmacovigilance awareness among health information managers in kwara state; ii. find out the perception of health information managers on management and utilization of covid-19 vaccine information; iii. ascertain the information management strategies available to manage information about covid-19 vaccines in kwara state; iv. identify the benefits health information managers derive from covid-19 vaccine information; v. identifies challenges in the management of covid-19 vaccines information. 2. material and method the study adopted a narrative based design. narrative based design is a way of conducting research in the qualitative mode. the qualitative mode of inquiry involves focusing and studying one or two individuals, gathering data by collecting their stories, reporting their individual experiences and chronologically organizing the meaning of the experiences collated (czarniawkska, 2004). the study’s population consists of eight purposively selected health information managers in various departments in kwara state ministry of health. these information managers are health officials in charge of covid-19 vaccine information management in the state. this category of health information managers was chosen because they have cogent and vital information on the theme of the study, based on the office they occupy. therefore, health information managers formed units of sample for the study while interview was used to gather data on their experiences. the interview schedule was considered appropriate for data collection health officials. the researchers personally visited the ministry of health to conduct face to face interview and the data collected were transcribed, coded and analyzed using narrative based technique. all ethical procedures for the research were ensured. https://doi.org/10.34010/injiiscom.v4i1.9572 67 | international journal of informatics information system and computer engineering 4(1) (2023) 61-74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 3. results and discussion the findings of the study based on the responses from the interview conducted on the eight health information managers in the study location was transcribed and presented based on the objectives of the study. 3.1. pharmacovigilance awareness among health information managers the responses of participants indicate their high level of awareness of pharmacovigilance and good knowledge of what the concept entails. one of the participants considers pharmacovigilance as a regulation, which assists in reaction to administration of drugs and vaccines. according to another participant, “pharmacovigilance is a practice that helps the monitoring of the adverse effects of drugs or vaccines and in evaluating reports to promote drug safety”. respondent 9 further described the concept as a process of continuous scrutiny, monitoring and evaluation of the drugs, vaccines medicine usage. other participants rationalized the concept as a ‘surveillance method’, ‘practice of monitoring’ and ‘sharing information’ to guide against the unseemly and side effect of licensed drugs or other pharmaceutical products. these findings indicate that pharmcovigilance is not only a concept but also an established regulation and practice to ensure safety of drugs and pharmaceutical products application and usage. 3.2. perception of health information managers on management and utilization of covid-19 vaccine information the results of the study on the perception of health information managers on the management and utilization of covid19 vaccines reveal several opinions from both administrators and users of the vaccines. for instance, one of the study participants reported high level of noncompliance to covid-19 vaccine usage instruction by patients and inadequate health personnel for quick intervention and administration of the vaccine. the situation was said to have led to the loss of many lives. another participant responded that community members refuse to come back for the second dose as a result of reaction to the first dose of the vaccine. this corroborated with hotez, nuzhath and colwell (2020) that vaccine hesitancy and refusal have become an important problem in recent years. this hesitancy against been fully vaccinated could be adduced to lack or low management of covid-19 vaccine information, and inadequate sensitization of community members on usage instructions. also, other study participants expressed fear of side effects and misconception of the purpose and usefulness of the vaccine caused by individual sensitivity to false information about the vaccine and rumors about aftereffects. this is in line with adejumo et al. (2021) findings on the perceptions of the covid-19 vaccine and willingness to receive vaccination among health workers in nigeria. adejumo reported that slightly more than half of the population of the health workers had positive perceptions of the covid-19 vaccine and unwilling to receive vaccination. furthermore, the respondents expressed varying opinions on how information about covid-19 vaccines is managed in the state. their responses show that sensitization is ongoing in the state through various medium such as the social media, fliers, bill boards, seminars https://doi.org/10.34010/injiiscom.v4i1.9572 salman et al., pharmacovigilance strategy for covid-19 vaccine information…| 68 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 and trainings. though most of the respondents maintained that information about covid-19 is strictly confidential, as such, much information on its management were conceived and reserved for health professionals. however, the respondents submitted that more sensitization is needed on covid19 vaccine. one of the respondents also noted that the sensitization should be on a weekly basis and address how the vaccine’s side effects can be prevented or attended to. the findings also revealed that database on screening, treatment and management of the vaccine procurement, administration and usage are maintained in the state. besides, it was noted that all the covid-19 protocols are observed when taking and administering the first and second doses of the covid-19 vaccine. also, it was revealed through the findings that demand for the vaccine increased with proper mobilization and awareness. the respondents noted that the vaccine and its accompanying information are kept in the state ‘cold store’ where the team named kwasphcda oversees the activities. 3.3. availability of information management strategies to manage covid-19 vaccines in information the results of the study reveal several approaches used by information managers to manage covid-19 vaccine information in kwara state. from the results, response of respondent 1, 2, 3 and 6 revealed that sensitization is done through the media to pass information about the reaction of any vaccine introduced to members of the community. in addition, the vaccine consumers are encouraged to report adverse reactions while effective documentation of reports is ensured. respondent 5 indicated early detection of adverse event of drugs (including vaccine) by questioning and monitoring as the information management strategy in use. the findings imply that sensitization on different media, report documented by the immune surveillance system, and records taken from reports and monitoring team is regarded is a major pharmacovigilance strategy available for covid-19 vaccine information management in the state. though a few of the respondent could not identify which strategy is being used for pharmacovigilance, some identified leaflets, posts, frequently asked questions, availability of cards (for data capture), e-registration, manual registration (register), and healthcare personnel in case of adverse events as alternative information strategies used. 3.4. benefits of covid-19 vaccine information to health information managers from the study findings several benefits of covid-19 vaccine information were identified by the health information managers interviewed. the respondents all agreed that covid-19 vaccine provided protection for health workers by boosting their immunity and creating opportunity for physical and virtual meeting without fear of getting infected by the virus. they likewise asserted that the vaccine has helped in guaranteeing the wellbeing of their loved ones (family, friends and relatives), updating their knowledge about preventive measures during pandemic. some other benefits identified by the participants include: provision of efficient management of scarce resources such as antiviral drugs and disease protective tools covid-19 https://doi.org/10.34010/injiiscom.v4i1.9572 69 | international journal of informatics information system and computer engineering 4(1) (2023) 61-74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 kit, free and fast access to the vaccine as it can be administered at a nearest facility to community members, reduction in mortality rate, reduction in unnecessary retainment of patients in the hospital, protection from contracting the virus while carrying out their duties as health officials, creation of more job opportunities for health workers and improved economy. moreover, the covid-19 pandemic and information on the vaccine have provided juicier opportunities for health care providers. the study findings on opportunities provided by the covid-19 vaccine in the state revealed that health care providers enjoyed free face masks and hand sanitizer, alongside some covid-19 containment tools. the result of the interview shows that these tools were also distributed free to the members of the public by the government. one of the study participants noted that the opportunity for meetings on virtual platform to reduce cost and stress of physical meeting, as well as trainings on emergency preparedness and response. some other opportunities included: increased global hygiene, research and vaccine production and opportunity to strengthen the health system. 3.5. challenges with managing covid-19 vaccine information the study results on challenges faced in the management of covid-19 vaccines’ information in the state. some participants referenced that “certain individuals perceived that when you take the vaccine, after 2years you will die’. likewise, they reported that people decline from been vaccinated due to some erroneous beliefs and rumors such as ‘after taking the covid-19 vaccine, the blood in the whole system will clot and will result to death immediately or any time soon’. other challenges identified include that of lack of acceptance, mistrust and unbelief in the government. some of participants reported that “some of the community members do not believe covid-19 is real”, “some believe it is rich-men disease”, “some others believe that the government is using covid-19 pandemic to make money”, “some likewise figure that the vaccine can kill” and “many saw it as a population reduction strategy, religious conviction or myth”. hence, masses do not trust government intervention, thereby making the work of health information managers and workers quite difficult. 4. discussion the study indicates high level of awareness of pharmacovigilance among health information managers. the findings of oshikoya and awobusuyi (2009) and fadare, enwere, afolabi, chedi and musa (2011) contradicts the current study opposes on lack of awareness on the availability and existence of national pharmacovigilance among healthcare providers in northern nigeria. this indicates much improvement in the level of awareness of pharmacovigilance in nigeria. opadeyi, fourrier-réglat and isah’s (2018) investigation in south-south nigeria established an increasing level of awareness of pharmacovigilance. similarly, adisa and omitogun (2019) confirmed that about three quarters of the health workers in ibadan, nigeria, were already aware of pharmacovigilance. this finding also corroborates the description of pharmacovigilance as a means of detecting, evaluating and preventing undesirable effects of drugs as explained by al-mustansir (2013), akarowhe (2020) and tyagi (2021). https://doi.org/10.34010/injiiscom.v4i1.9572 salman et al., pharmacovigilance strategy for covid-19 vaccine information…| 70 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 likewise, the study findings agree with santosh and tragulpiankit (2011) that pharmacovigilance is an ongoing process to monitor drug safety and to make available new information and knowledge about adverse drugs reactions. it can be deduced from the results that there is a positive increase in the level of awareness of pharmacovigilace anomg health managers. the study therefore reveals pharmacovigilance as a usual and beneficial practice among health information managers in the state. it can also be deduced from the findings on the perception of health information managers on the management and utilization of covid-19 vaccines, that effective national implementation of covid-19 vaccination depends on the ability to access, receive and share information rapidly between health information managers and staff at all levels. in order to facilitate informed decision-making from diverse sources related to covid-19, such as local epidemiology, vaccination readiness, performance and coverage tracking, vaccine and supply management, vaccine safety, and immunization waste management, should be gathered, analyzed, and disseminated effectively. real-time monitoring of covid-19 vaccine uptake, safety, and effectiveness could be enhanced by employing electronic information platforms, although paper-based or partially digital systems may also be suitable. the results on availability of strategies employed in managing covid-19 vaccine information by health information managers in kwara state indicate the multifaceted approaches used. some of these approaches have generated diverging opinions about the management of the vaccine. this type of behavior may be due to inappropriate information management strategy and largely fall into the category declared by aronson (2009) as therapeutic errors which require that pharmacovigilance stakeholders reactivate risk minimization strategies. along with the ability to organize a system of emergency medical care, the communicative effectiveness of government has come to the fore during the current pandemic. in this regard, it seems relevant to use the concept of “communicative efficiency,” relating to information management strategy as risk minimization strategies. crowley et al. (2021) utilized the theory of motivated information management (tmim) to examine the associations between uncertainty about covid-19 illness and information management. the results revealed that perceived uncertainty, anxiety, and strategies for managing information were significant predictors of adopting covid-19 preparation and prevention behaviors. several studies also support that information on covid-19 vaccine is beneficial to health workers as well as individuals in a community. hoge (2021) highlighted the benefits of covid-19 vaccines to include: opportunity to organize and attend small events in person, travelling safely without fear of infections, reconnection with family, loved ones and co-workers, in addition to the immediate health and long term benefits. it can be stated expressively that health workers are largely benefitting from the vaccine information, seeing they usually are at the front-line when there is an outbreak. thus, opportunities and benefits provided by covid-19 vaccine (unesco, 2021) calls for considering the vaccine as a general public good and https://doi.org/10.34010/injiiscom.v4i1.9572 71 | international journal of informatics information system and computer engineering 4(1) (2023) 61-74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 ensuring its ideal, affordable and timely access. on the results of challenges faced in the management of covid-19 vaccines’ information, the erroneously passed information about the vaccine seem to have formed a belief in people’s consciousness, subsequently denying further penetration of significance and advantages of the vaccines in the individual’s minds. since there is probably no successful exercise an organization can undertake with surviving some challenges. thus, efficient data sharing strategies must be ensured in managing covid-19 information. some of the study participants indicate the need for more sensitization in order to disabuse the public’s negative perception about the covid-19 vaccination. in a study conducted by gao et al. (2020) on the management and sharing of information during the covid-19 pandemic, it was found that data management systems are encountering numerous challenges and issues. these challenges include data decentralization, inconsistencies, security and legal concerns, limited financial support, international communication, standardization, and globalization. they accordingly suggested the development of several integrated platforms or models for national and worldwide data-sharing and management in order to overcome and resolve the identified problems. the health information management system if well designed will help to meet the unprecedented need and demand for covid-19 vaccine information sharing and research worldwide. 4.1. summary of findings the study reveals that majority of the respondents are well aware and have knowledge of pharmacovigilance as it pertains to their organization and the state understudied. furthermore, the findings declare expression of fear on rumours on side effects and misconception of the purpose and usefulness of the vaccine as factors affecting management and utilization of covid-19 vaccine information. however, the vaccines’ information sources are managed through who guidelines on covid-19 vaccines. though several information management strategies are available, however, there strategies are diverse including sensitization on different media, report documentation using surveillance system, monitoring team, and healthcare personnel in case of adverse events. the development of covid-19 vaccines has also created various benefits and opportunities for health information managers; including: immunity boost, virtual meetings, access to scarce of health resources, prevention of severe illness and reduction of hospitalization, trainings on emergency readiness and response, improvement in personal hygiene and research, reinforced national health system and protection for family and friends and the community. the study identified vaccines hesitancy, refusal, noncompliance to immunizations, fear of unknown, false reports, misinformation, overwhelmed health facilities and government’s deceit, as some of the challenges affecting the management of covid-19 vaccines information. 5. conclusion the prevailing demand to carry on research as part of the continuing allinclusive response to covid-19 vaccine has driven the conduct of this small-scale population-based prevalence survey particularly in kwara state in order to https://doi.org/10.34010/injiiscom.v4i1.9572 salman et al., pharmacovigilance strategy for covid-19 vaccine information…| 72 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 determine the prevalence covid-19 vaccine information management strategy and utilization in the state. though the health information managers in the state are well informed of pharmacovigilance and its importance, they however are affected by rumours and fear of side effects and misconception of the populace on the usefulness of the vaccine. thus, factors affecting how covid-19 vaccine information is being managed. as such, the complex but very diverging strategies used in response to proper management have given way for alternatives which are misleading. the study findings call for an urgent intervention in the level of sensitization of community members, as well as health information managers on benefits of the vaccine. this can be specially achieved at the primary health care level due to their closeness to the community. the restricted precautionary choices in the state can be enhanced with the development of health surveillance and information management systems, and the reformation of pharmacovigilance control. such a survey could be a platform for generating reliable data on the epidemiology of covid-19 vaccine of public health importance. procurement of the vaccine’s variant is more in consistence with our realities if we are going to achieve a nationwide coverage. finally, health information managers in charge of vaccine in kwara state are in position to effectively and efficiently maximize vaccination campaign and positively utilized information against negative covid-19 misinformation in the state culminating in eradication of covid-19 in nigeria. the provision of advance information management technology is hereby recommended to enhance pharmacovigilance among health information managers in state. also, there ought to be harmony in the information management strategies available, so they can serve as a guide for information managers. information about how covid-19 vaccines should further be managed in cognizance with who guidelines of covid-19 vaccines. similarly, the benefits derived from health information managers through covid-19 vaccine should be advocated through sensitization and campaigns to increase level of awareness. an independent integrated platform model for data-sharing and management would help to overcome the difficulties in the management of covid-19 vaccines’ information in the country. references abdel-latif, m. m., & abdel-wahab, b. a. (2015). knowledge and awareness of adverse drug reactions and pharmacovigilance practices among healthcare professionals in al-madinah al-munawwarah, kingdom of saudi arabia. saudi pharmaceutical journal, 23(2), 154-161. abdullateef, m., & okonkwo, b. (2021). why trusting nigeria centre for disease control information is difficult: a study of the awareness, knowledge and perceptions of nigerians hesitant towards covid-19 management by ncdc in nigeria. traektoriâ nauki= path of science, 7(5), 6001-6008. adejumo, o. a., ogundele, o. a., madubuko, c. r., oluwafemi, r. o., okoye, o. c., okonkwo, k. c., ... & ohaju-obodo, j. o. (2021). perceptions of the covidhttps://doi.org/10.34010/injiiscom.v4i1.9572 73 | international journal of informatics information system and computer engineering 4(1) (2023) 61-74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 19 vaccine and willingness to receive vaccination among health workers in nigeria. osong public health and research perspectives, 12(4), 236. adisa, r., & omitogun, t. i. (2019). awareness, knowledge, attitude and practi ce of adverse drug reaction reporting among health workers and patients in selected primary healthcare centres in ibadan, southwestern nigeria. bmc health services research, 19, 1-14. aleinikov, a. v., kurochkin, a. v., & maltseva, d. a. (2021). informatio n management under the conditions of global pandemic. scientific and technical information processing, 48, 38-44. cucinotta, d., & vanelli, m. (2020). who declares covid-19 a pandemic. acta bio medica: atenei parmensis, 91(1), 157. czarniawkska, j. (2004). narrative base design. london: facet publishing. dal pan, g. j. (2014). ongoing challenges in pharmacovigilance. drug safety, 37, 1-8. enger, c., younus, m., petronis, k. r., mo, j., gately, r., & seeger, j. d. (2013). the effectiveness of varenicline medication guide for conveying safety information to patients: a rems assessment survey. pharmacoepidemiology and drug safety, 22(7), 705-715. erameh, n. i., & ojakorotu, v. (2021). the nigerian state, corruption and the political economy of covid-19 governance. gender and behaviour, 19(1), 17597-17608. fadare, j. o., enwere, o. o., afolabi, a. o., chedi, b. a. z., & musa, a. (2011). knowledge, attitude and practice of adverse drug reaction reporting among healthcare workers in a tertiary centre in northern nigeria. tropical journal of pharmaceutical research, 10(3). gao, f., tao, l., huang, y., & shu, z. (2020). management and data sharing of covid19 pandemic information. biopreservation and biobanking, 18(6), 570-580. germani, f., & biller-andorno, n. (2021). the anti-vaccination infodemic on social media: a behavioral analysis. plos one, 16(3), e0247642. hoge, b. (2021). five benefits of getting a covid-19 vaccine. current covid-19 health and safety guidelines. [online] https://www.uab.edu/news/youcanuse/item/12025-five-benefits-ofgetting-a-covid-19-vaccine (accessed 16 november 2021). hotez, p. j., nuzhath, t., & colwell, b. (2020). combating vaccine hesitancy and other 21st century social determinants in the global fight against measles. current opinion in virology, 41, 1-7. hussain, a., tahir, a., hussain, z., sheikh, z., gogate, m., dashtipour, k., ... & sheikh, a. (2021). artificial intelligence–enabled analysis of public attitudes on facebook and twitter toward covid-19 vaccines in the united kingdom and the united states: observational study. journal of medical internet research, 23(4), e26627. mahase, e. (2020). covid-19: who declares pandemic because of “alarming levels” of spread, severity, and inaction. bmj, 368(march), m1036. maxmen, a. (2019). this nigerian doctor might just prevent the next deadly pandemic. nature, 566(7744), 310-314. https://doi.org/10.34010/injiiscom.v4i1.9572 https://www.uab.edu/news/youcanuse/item/12025-five-benefits-of-getting-a-covid-19-vaccine https://www.uab.edu/news/youcanuse/item/12025-five-benefits-of-getting-a-covid-19-vaccine salman et al., pharmacovigilance strategy for covid-19 vaccine information…| 74 doi: https://doi.org/10.34010/injiiscom.v4i1.9572 p-issn 2810-0670 e-issn 2775-5584 national academy of sciences and national academy of medicine (2021). nas and nam presidents alarmed by political interference in science amid pandemic. [online] https://www.nationalacademies.org/news/2020/09/nas-andnampresidents-alarmed-by-politicalinterference-in-science-amid-pandemic opadeyi, a. o., fourrier-réglat, a., & isah, a. o. (2018). assessment of the state of pharmacovigilance in the south-south zone of nigeria using who pharmacovigilance indicators. bmc pharmacology and toxicology, 19, 1-8. scott, s. (2021). what is health information management? [online]. https://www.herzing.edu/blog/what-health-information-management (accessed 29 october, 2021). tyagi, s. (2022). global research output in ‘pharmacovigilance’during 2010 – 2020. therapies, 77(3), 273-290. unesco (2021). covid-19 vaccines: challenges, weaknesses and opportunities in the region. montevideo. [online]. https://en.unesco.org/news/covid19vaccines-challenges-weaknesses-and-opportunities-region. vallano, a., cereza, g., pedròs, c., agustí, a., danés, i., aguilera, c., & arnau, j. m. (2005). obstacles and solutions for spontaneous reporting of adverse drug reactions in the hospital. british journal of clinical pharmacology, 60(6), 653-658. wang, w. t., & wu, s. y. (2021). knowledge management based on information technology in response to covid-19 crisis. knowledge management research & practice, 19(4), 468-474. williams, s. n., & dienes, k. (2021). public attitudes to covid-19 vaccines: a qualitative study. medrxiv, 2021-05. yamey, g., & gonsalves, g. (2020). donald trump: a political determinant of covid 19. bmj, 369. https://doi.org/10.34010/injiiscom.v4i1.9572 https://en.unesco.org/news/covid19-vaccines-challenges-weaknesses-and-opportunities-region https://en.unesco.org/news/covid19-vaccines-challenges-weaknesses-and-opportunities-region 105 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 a simple, low-cost, efficient and smart consumer gas leakage detection system benjamin kommey*, daniel opoku, jessica lartey, kirston kumah, mark tantuo kwame nkrumah university of science and technology, kumasi, ghana *corresponding email: bkommey.coe@knust.edu.gh a b s t r a c t s a r t i c l e i n f o liquefied petroleum gas is used for a wide range of both domestic and industrial purposes. its use has seen a significant increase over the years. unfortunately, there have also been many casualties resulting from gas explosions in recent times. often, there is a news report of a gas explosion occurring somewhere, whether at home, in the marketplace or in an industry. although these accidental explosions are sometimes beyond human control, some are due to ignorance, negligence, or improper care of gas cylinders. most explosions are caused by undetected gas leakages which are then set off to explode with the slightest ignition. apart from explosions, fires at homes can also be caused by gas leakages. this usually results in serious injuries, loss of property and even sometimes, loss of lives. this project seeks to provide a smart, budget friendly and easy to use gas leakage detector system, a better alternative to relying on the natural way of using the nose, which may fail and in effect, incur undesirable consequences. the system employs an mq2 sensor interfaced with an esp8266 wi-fi module for sensing the gas in the atmosphere and displaying an alert message on the lcd. the wi-fi in combination with the microcontroller module acts as the transmission and processing unit, allowing for communication over wireless medium. this way, a user can receive notifications on a mobile device depending on changes the hardware system detects in the atmosphere. there is also a buzzer to alert nearby persons by sound so that they can take the necessary precautions and actions. article history: received 25 may 2022 revised 30 may 2022 accepted 10 june 2022 available online 26 june 2022 aug 2018 __________________ keywords: lpg, gas cylinder, mq2-sensor, explosion, gas leakage international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 105-123 mailto:bkommey.coe@knust.edu.gh benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 106 1. introduction in may 2021, a survey conducted by doris dokua revealed that 267 fire outbreaks were registered in the greater accra region of ghana alone. domestic fires were the most prevalent source of outbreaks, having 108 occurrences, with electrical installations being the next at 48, and the others being either bush fires, vehicular fires, institutional fires, or fires from burning refuse and others (addai, et al., 2016). fire is an oxidation process that occurs rapidly and is usually accompanied by the evolution of heat and light in varying intensities. four components are thought to be necessary for fire to start: fuel, an igniting source, an oxidizing agent (typically ambient oxygen), and a reaction mechanism. the most frequent causes of flames include flammable gases used for cooking in houses, furniture, clothing, solvents like kerosene and gasoline used in daily activities, and combustible dusts like toner used in offices and starch sold in markets (doris, 2020). liquefied petroleum gas, lpg, is a flammable mixture of hydrocarbon gases, most commonly propane, butane, and propylene (raslavičius et al., 2014). it is a colourless and doorless gas widely used in the domestic setting for a variety of purposes, mostly being cooking, and other times, as fuel for vehicles. to allow portability and mobility, lpg must be placed under pressure to be transported as a liquid, since it occupies much less volume as a liquid and is hence transported in gas cylinders. when this pressure is released, such as when the nozzle on a gas cylinder is open, the liquid becomes gas again (murshed et al., 2021). lpg has a wide array of uses, which vary from domestic to industrial. it is used as fuel for vehicles, it is used for cooking, used as a refrigerant, used as chemical feedstock, used for drying processes in agriculture and industrial solution heating, and is also used to produce electrical energy by running turbines (checkley et al., 2021). despite the benefits lpg provides, there are dreadful consequences if certain measurable quantities of the gas can be detected in the atmosphere. lpg serves the average person well usually as fuel but fails sometimes. according to a study conducted by the zebra in 2020, the top three causes of house fires are cooking, heating equipment and electrical malfunction, with cooking alone contributing to 50% of all fires. each year, there is an average of 358,300 home-based fires, causing nearly 12 billion dollars in damage (tylor, 2020). statistics from the ghana national fire service suggests that domestic fire outbreaks are peaking, with the highest in 2019 and 2020. the domestic fire incidents from january to october represented 39.63 per cent and 35.67 per cent of total fire outbreaks in both years respectively (news desk report, 2021). most house fires in ghana are caused either by lpg explosion or leakage. lpg is colourless and odourless. its nature makes it difficult to perceive by sight or by smell alone in the case of a leakage which poses danger to everyday users, hence the need for a monitoring device to reduce the risk of harm. this paper seeks to prevent the occurrences of fires and explosions owing to undetected gas leakages by providing a smart gas 107 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 detection system for consumers. this will be done by incorporating software and hardware components to build a gas leakage detector which is cost effective and efficient in detection of gas leakages and the presence of fires. the system is also efficient with alerting users in case of gas or flame detection and is integrated with a mobile application for constant monitoring of the hardware system. 2. related works lpg consists of a mixture of propane and butane produced by walter snelling in the year 1910. this gas is a highly flammable chemical compound that is odourless and is heavier than air (noku et al., 2021). the undetectable yet flammable nature of this gas makes it very dangerous to accommodate in the home; yet it is very common in many ghanaian homes and industries serving various purposes because of its economic benefits. an odorant such as ethane oil is added to allow for natural detection by the average user, yet this gas is still not easily detectable by many at certain concentrations which renders it unreliable. with regular use, gas cylinders and pipes reach a point of wear and tear, leaving the gas prone to leakage, which is why these systems require constant monitoring. n. evalina and azis h designed a system consisting of the atmega 8 microcontroller, an mq 6 sensor and a buzzer. it was a simple, cost-effective system designed to detect the leakage of lpg gas and to alert people by the sound of the loud buzzer (evalina & azis, 2020). noku et al proposed a design of an automatic gas leakage detector with a control system and auto power trip. the control system included a gas valve for cutting supply of gas and was also responsible for shutting down electrical systems to avoid sparks which could lead to a fire. k. r. katole et al in 2019; proposed a system that employed the mq 4 sensor for methane gas detection, the mq 2 sensor for lpg leakage detection, and the mq 7 sensor for the detection of carbon monoxide. this system generally performed gas leakage detection and alerted people by displaying a message on an lcd screen stating the specific gas detected. in any case of gas leakage, people had to rely on their nose if they were far from the lcd and could not read what was displayed which is not the best safety mechanism applicable in this area (katole et al., 2019). e. jebamalar et al worked on a simple system which involved the use of the mq 6 sensor interfaced with a microcontroller which generally performed gas leakage detection, and then alerted people around by activating the buzzer so that they could take the necessary precautions (leavline et al., 2017). syeda bushra shahewaz et al proposed a system using arduino uno. this system employed three different gas leakage detectors planted at different parts of a building to cover more area. these gas leakage detectors were all linked to the same microcontroller and displayed messages on an lcd screen when gas is detected in a particular zone, specifying where. it also used a buzzer and a gsm module to alert people by sound and by sms respectively (shahewaz & prasad, 2020). hitendra et al designed a system that involved the use of a pic microcontroller and an mq 6 sensor. it was basically responsible for detecting benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 108 gas leakages and for sensing extreme temperatures which gives a clue that there may be a fire (rawat et al., 2014). a. y. nasir et al presented an alternative approach to develop a device that can automatically detect and control gas leakages and monitor temperature in vulnerable areas. the system detects the leakage of lpg using a gas sensor and then also monitors the temperature using a temperature sensor. when the lpg concentration in the air exceeds a certain level, the gas sensor senses the gas leakage and the output of the sensor goes low, the system then opens the exit windows, and then uses a gsm module for an alert about the gas leakage. also, when the temperature of the environment exceeds a certain limit, it then turns on the led (indicator) and makes an alarm sound through the buzzer. an lcd displays the current temperature and gas leakage status in degree celsius and ppm respectively (nasir et al., 2019). rhonnel s. paculanan and israel carino proposed a system to detect the presence of lpg leakage as a part of a safety system. apart from alarm by sound, an sms alert would be sent to inform the authorized person of the leakage and the solenoid valve of the gas container will be triggered to shut down the gas supply to prevent any harmful effects due to gas leakage. essentially, a gas sensor was used to monitor the lpg if the gas leak reached beyond the normal level. this proposed leak detection system would then trigger the sound alarm. there was also a survey conducted on 75 people to know whether most people who use lpg are household members or employees of companies (paculanan & carino, 2019). bader farhan alshammari and muhammad tajammal chughtai in their paper outlined an industrial monitoring system design using the internet of things (iot). here, the gas sensor (mq-5) captured information which is posted into a data cloud. the sensor detects the leakage of gas under most atmospheric conditions. all the components are controlled by an arduino (uno-1) that acts as a central processor unit in the setup. as soon as a gas leakage is detected by the sensor, the alarm is raised using a buzzer. this alarm is supported by an lcd to display the location of leakage, alert the observer, and automatically activate the exhaust fan in the particular section to extract leaked gas from the enclosed space to prevent potential ignition and fire hazards (alshammari & chughtai, 2020). in a paper by mohammad monirujjaman khan, an mq6 semiconductor sensor is used to detect lpg gas. sensitive material of the mq-6 gas sensor is sno2, which has lower conductivity in clean air. when the target combustible gas exists, the sensor conductivity increases along with the rising gas concentration. the mq6 gas sensor has a high sensitivity to propane, butane and lpg, and response to natural gas. the sensor could be used to detect different combustible gasses, especially methane; it has a low cost and is suitable for different applications. the mq-6 can detect gas concentrations anywhere from 200 to 10,000 ppm. the sensor’s output is an analog resistance. when the sensor detects gas in the atmosphere, it will give digital output 1 and if gas is not detected 109 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 the sensor will give digital output 0. arduino will receive the sensor output as digital input. if the sensor output is high, then the buzzer will start tuning along with the lcd that will show that “gas detected: yes”. if the sensor output is low then buzzer will not be tuning, and the lcd will show that “gas detected: no”. the buzzer consists of several switches or sensors connected to control unit that determines which button was pushed or whether a preset time has lapsed, and usually illuminates a light on the appreciate button or control panel and sounds a warning in the form of a continuous or intermittent buzzing or beeping sound (khan, 2020). movasudev yadav in his project sought to add a little twist to the usual gas detection module by introducing a stepper motor also in addition to the normal lpg gas leakage detectors which helps in turning off the switch when there is an emergency in our absence. the sensor used here has excellent sensitivity combined with a quick response time. the sensor can also sense iso butane, propane, lng, and cigarette smoke. the report also outlines a background into the area of 8051 microcontroller and mobile communication, how they are interfaced to each other, and at commands set used in communication (yadav et al., 2016). ba thanh nguyen, anh vu nguyen designed a gas leakage detection and monitoring system which included an arduino mega 2560, an mq-2 sensor, a sim800l module, an esp8266 node mcu module, an lm2596 module, a lcd, and a buzzer. an allowable limit of gas sensed in the atmosphere was set. if the gas level exceeds 20% of the allowable limit, the system is triggered, data is displayed on the lcd showing that gas leakage has been detected, and the sim alarm clock calls the phone number set before the gas exceeds the limit. data is also posted to thingsspeak.com every 15 seconds to ensure constant gas monitoring (nguyen & nguyen, 2020). rahul gupta et al designed a system with an at mega 328 microcontroller interfaced with two stepper motor ic (uln 2003a), buzzer, lcd (liquid crystal display), gsm module and rf link. upon gas leakage detection, the system is triggered, and the buzzer activates simultaneously with a message displayed on liquid crystal display screen, and the gsm module is activated, which sends warning sms to the user. the stepper motor ic (uln 2003a) drives the stepper motor attached to it causing the main power to go off (gupta, 2018). arun raj et al, designed a pic 16f877a microcontroller with an operating frequency of 20 mhz. it was interfaced with a gas sensor, weight sensor (load celll6d), gsm module (simcom 300), and display(s). it was a simple-to-use lpg sensor, suitable for sensing lpg concentrations in the air. the load cell acts as a passive transducer or sensor for converting applied force into electrical signals. this was used to check the available gas level. this system gives a fully automated approach towards the gas booking along with gas leakage detection. real time weight measurement of the gas and its display on lcd makes it an efficient home security system (raj & athul, 2015). designed a gas leakage detection system interfacing an at mega 328p microcontroller with an mq 6 gas sensor, an lcd, a stepper motor, and a buzzer benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 110 (pooja, 2020). this functioned as an effective gas leakage monitoring system which simultaneously displayed a message on lcd while sending a message to a specific mobile number using the gsm and activating the buzzer to sound an alarm when gas was detected in the atmosphere. there was also a reset button available for resetting the system to its former state after the gas leakage has been fixed (imade et al., 2018) used the iot technology to make a gas leakage detector for society which will have smart alerting techniques involving sending text messages to the concerned authority and an ability to perform data analytics on sensor readings. the primary objective was to propose the gas leakage system for society where each flat has a gas leakage detector hardware. this will detect the harmful gases in the environment and send an alert to the society member through alarm and sending notification. proposed a solution to this problem by building a device utilizing sensors connected to the node mcu (sudha & prasad, 2020). the device performs area monitoring continuously. the gas sensor provides data to node mcu, and then the results are displayed as a warning to the user via an androidbased smart-phone device. other than lpg gas, air conditioner and refrigerator leaked gases are also harmful in the home. using this device, users will be able to prevent accidents that occur due to harmful gas leaks so that accidents can be avoided. proposed an internet of things (iot) based system that maintains track of several lpg cylinder-related factors and updates the user via a mobile application (rohith, 2020). there are two sections to this project. the first part deals with the amount of gas still in the cylinder, which is calculated by a load cell sensor and continuously updated to the user's smartphone. an alert will be delivered to the user via the buzzer and a notification in the mobile app when the gas level is below the threshold value. another component is the safety feature built into this home automation system, which notifies the user of a gas leak by both a buzzer and a mobile application when it is discovered. it alerts the user of a gas leak via a buzzer and a mobile application when it is found. additionally, it turns off the gas regulator knob to halt additional leaks at their source. the gas valve can be turned on and off using the mobile application. the device is designed to serve as a stand for lpg cylinders. 3. method the method used for this is multi-faceted. the system is made up of several functional units that work together to provide the desired outcomes. these hardware and software components are incorporated into the system to carry out the smart gas leak detection functions. the system is visually represented here in the form of a system architecture, which depicts the system's expected functionality in general. along with a flowchart outlining the logic of the system, it also includes a block diagram that illustrates how the hardware and software components work. 111 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 3.1. system architecture the general operating principles of the smart gas leakage detector are outlined in the system architecture in fig. 1. fig. 1. system architecture it provides a visual representation of how the system works, outlining the various elements that must be recognized in the atmosphere, how the system handles detection, and how the user interacts with the system. the gas detector represents the whole system with the underlying components including sensors, a processor, alarms, and a wi-fi communication module. these components together makeup the hardware components of the system, collecting data from the environment and responding appropriately, by setting off an alarm and control system, and alerting the user through the mobile application. data collected from the atmosphere by the sensors is constantly uploaded via wireless media to a website, thingspeak.com, which the mobile software application depends on for updates about the state of the system. the mobile software application can also be used to remotely control the hardware components of the system. 3.2. system block diagram the smart gas detection system consists of three major units; the sensor unit, the processing unit and the alarm unit as depicted in fig. 2a and 2b. as seen in figures, the sensor unit consists of components including the gas and smoke sensor as well as the flame sensor. these components are constantly on the lookout for changes in the atmosphere with respect to smoke, flames, and lpg concentrations. the processing unit is the intelligent unit which consists of a processor which receives and interprets digital signals from the sensor unit. based on what data is received, signals may be sent from this unit to trigger the alarm unit to alert people nearby to take precautions. the alarm and response unit are made up of the lcd, the led, the buzzer, the sprinkler, and the solenoid valve. by transmitting signals that notify the user of the detection of gas in the environment as well as the risk of a fire occurring, these hardware components work together to engage the user's corporeal senses and enable the situation to be mitigated. (a) (b) fig. 2. system block diagrams benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 112 3.3. system workflow the system workflow shows the sequential stages that the gas sensor traverses to achieve its function of gas detection and alarming the user. the individual phases are as illustrated in fig. 3, giving a graphical perspective of the logical flow of operations in the smart gas detector system. the system starts with the press of a button, which turns on the green led after which the sensors begin to compare atmospheric concentrations of gas with previously defined values. in case the values are detected to be higher than allowable concentrations, the solenoid valve is closed to obstruct the supply of gas while the red led is turned on. the buzzer is also turned on to sound an alarm, while the lcd displays a message indicating either gas detection or flame detection. if after a while, the concentrations as previously detected reduce to normal allowable levels, the system then resets and begins the process all over again. notifications are also sent to a user via the system mobile software application. fig. 3. system workflow 3.4. system hardware prototype design the system hardware prototype was designed using modular method. the modules or hardware prototype components are illustrated in fig. 4 and graphically and textually described in detail in table 1. fig. 4. system hardware prototype design 113 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 table 1. system hardware prototype components and description components image description mq-2 gas sensor gas sensors are devices that can detect the presence and concentration of various hazardous gases and vapours, such as toxic or explosive gases, volatile organic compounds (vocs), humidity and odours (mirzaei et al., 2018). a gas sensor uses specialized physical and chemical processes to transform the constituent parts and concentrations of various gases into conventional electrical signals. the mq-2 sensor is sensitive to smoke, as well as other flammable gases including lpg, butane, propane, methane, alcohol, and hydrogen. it can be set to output values either in the form of analogue signals or digital signals which can be read with the analogue or digital inputs of the arduino respectively. flame sensor a flame detector is a sensor designed to detect and respond to the presence of a flame allowing flame detection. the flame detector can often respond faster and more accurately than a smoke and heat detector due to the mechanisms it uses to detect the flame. this sensor is a short length of a thin metallic rod that creates a small current of electricity to confirm there is a fire burning. it can detect infrared light with wavelengths between 700nm and 1000nm. the far-infrared flame probe converts the infrared light detected into current changes. with a detection angle of 60 degrees, the integrated variable resistor allows for sensitivity adjustment. it has a working voltage ranging from 3.3 to 5.2 volts dc, and a digital output used to show whether a signal is present (https://create.arduino.cc). node mcu esp8266 nodemcu is an easy-to-use open-source platform built on the esp8266 which is programmable with arduino ide or iua languages. it allows things to be connected and data to be transferred over wifi. additionally, it may address many of the project's requirements on its own by offering some benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 114 of the most crucial microcontroller functionalities, like gpio, pwm, adc, and others, having an internal antenna, and containing 13 gpio pins, 10 pwm channels, i2c, spi, adc, uart, and 1-wire. lcd a liquid crystal display is a flat panel display or other electronically modulated optical device that uses the light modulating properties of liquid crystals combined with polarizers. liquid crystals do not emit light directly, instead using a backlight or reflector to produce images in colour or monochrome (liu & abbatt, 2021). the lcd in this system displays messages to alert the user about the status of the system. led a light emitting diode is a semiconductor diode which emits light when a voltage is applied to it. the leds in this design are responsible for representing the status of the system. the green led lights up when the system is turned on, while the red one only lights up when the system detects either gas, smoke, or flames. buzzer a buzzer or beeper is a mechanical, electrochemical, or piezoelectric audio signaling device. types of buzzers and beepers include alarm devices, timers, train, and confirmation of user input such as a mouse click or a keystroke (siregar & qurniawati, 2022; gustov & levina, 2021). the buzzer comes as an extra component for sending alarm signals to people within hearing distance. sprinkler a fire sprinkler system is a straightforward yet essential active part of the fire protection system in a residential or commercial building. the system, which is installed inside walls or ceilings, consists of sprinkler heads, a water source, and a pipe system for water distribution. the sprinkler automatically releases water when a fire is detected; smoke does not cause it to do so. a fire detection and suppression system both detects and suppresses the fire. sprinkler systems have 115 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 3.5. system mobile software application the gas detection system can be remotely observed at any time and from any location by a user via the mobile application. the requirements for the mobile software application were influenced by the project's objectives. through the smartphone application, a connection between the hardware system and the user is made possible. this section covers the functional and nonfunctional requirements for the mobile software application. it also provides a use case and activity diagrams, as well as a few application views. some of the development tools used are discussed as well. 3.5.1. software system functional requirements the functional requirement of the mobile software application provides a thorough explanation of what it should be able to achieve. the user needs and other requirements that the user may not directly interact with are presented in this section. the functional specifications for the mobile software application are listed. ● it should provide a login functionality to the user which authenticates the user when they allow the application to access email. been in use since hiram stevens maxim invented them in the late 19th century. they are currently widely utilized, with more than 40 million sprinkler heads installed annually. in most systems, water is kept under pressure in a system of pipes that runs throughout a structure. solenoid valve controller the definition of a solenoid valve is an electromechanical valve that is commonly used to control the flow of liquid or gas. in most flow control applications, it is necessary to start or stop the flow in the circuit to control the fluids in the system. an electronically operated solenoid valve is usually used for this purpose. by being solenoid actuated, solenoid valves can be positioned in remote locations and may be conveniently controlled by simple electrical switches. solenoid valves are the most frequently used control elements in fluidics. they are commonly used to shut off, release, dose, distribute or mix fluids. for that reason, they are found in many application areas. solenoids generally offer fast and safe switching, long service life, high reliability, low control power and compact design. benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 116 ● it should display a dashboard to a successfully logged in user. ● the dashboard shows the gas concentration and a progress circle updated in real-time indicating gas percentages in the atmosphere. ● it should have buttons for directly controlling the hardware components. ● it should display previous events on a history page. ● it should provide a logout functionality to allow users to logout of the application 3.5.2. system software non-functional requirements the features of the mobile software application are not specifically related to the non-functional requirements. nevertheless, they are important because they highlight how simple it is for users to interact with the software. the nonfunctional requirements for the mobile software application are: ● performance: this is a measure of how quickly the mobile application reacts to user input. the mobile application must be able to display measured data from the gas sensor in real-time with the least amount of lag time possible. ● security: as a measure to enforce security, only users who are logged in will be able to access the software application. ● ease of use: the user-friendly interface of the mobile software application makes using and navigating intuitive. it will be used to closely monitor readings from the gas sensor. 3.6. system use cases the system use cases are illustrated in the form of diagrams in fig.5. the use case diagram of the system provides a visual representation of the relationship that exists between actors internal or external entities that interact with the system and the system. the two actors in the system are the server, which provides real-time updates of measurements taken from the sensors in the gas detector. fig. 5. system use case diagram 3.7. system software activity diagram the system software activity diagram is as depicted in fig. 6. the activity diagram shows the possible activities of the mobile software application. when a registered user successfully logins in, the user is redirected to the home screen. if the user is not registered, they are redirected to the signup screen which allows a new user to get registered. from the home screen, a user can navigate either to the controls page or to the history page, and likewise from any of the other two pages. they can logout of the application at any time of convenience from any of the three screens. for an unsuccessful login, the user stays on the login screen so that they can try to login again. 117 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 fig. 6. system software activity diagram 4. results and discussion various tests were run to see if all functionalities were working as expected. table 2 contains screen shoots, functionalities, and test results. table 2. system software test and results task screen results signup unregistered users trying to login to the mobile application get redirected to the signup page. there, they are required to create an account using an email and a password. this allows them to access the functionality of the mobile application without obstruction. from here, they are redirected to the home page. login on the login page, the mobile application verifies if a user has been registered, the user is verified; if found, the user is directed to the home screen. verification is done with credentials used for registration on the signup page which include an email and a password. if an unregistered user tries to log in, they are redirected to the signup page where they can register with their preferred credentials. benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 118 home this screen is accessible only to authenticated users. the home page presents the user with the information they need about gas concentrations detected in the atmosphere, and if it is safe to be in that place. it represents gas concentrations visually using a progress circle and percentages. controls the controls page consists of four buttons phone, reset, solenoid valve and sprinkler. the phone button opens the user’s phone application on their mobile device with an emergency number already dialled in so that they can quickly place a call for help. the reset button is engaged when a user wants to reset the hardware system. the solenoid valve button sends a command to have the solenoid valve connected to the gas cylinder triggered so that it shuts off gas supply. the sprinkler button is used to manually activate the sprinkler connected to the hardware system in case it fails to start automatically. 119 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 history the history page displays data about previous states of the system, relating them to specific dates. this is important for referencing, as it is crucial for a user to closely monitor changes in the system. navigation the three dots at the upper right corner of the mobile application are responsible for navigation. engaging that button on a page shows a dropdown menu with buttons that lead to the two other screens. there is a logout button as well, which allows a user to end their session in the mobile application. repeated testing was done to expose situations that obstructed proper functioning of the system. during testing, it was ensured that test vectors were kept within perceivable range so that the sensors could detect them to give desired results. the gas detector, which includes both gas and fire sensors, was successful in logging data from those sensors to thingspeak.com to generate a visualization of past activities. table 3 contains a summary of various test vectors used during testing and how the system responded to them. in case the gas sensor detects gas in the atmosphere, the system must light a red led, sound the buzzer, display a message on an lcd and send a notification to the user’s mobile device. the same reaction was expected of the system when fire was detected, as well as when both gas and fire were detected. using the table, a success rate of 91.67% was obtained, since the system failed during trial 8. the success rate was determined using equation (1). success rate = (success / number of trials) x 100% (1) figs. 7 and 8 show the visualization of gas sensor and fire sensor data recorded over time on thingspeak.com. benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 120 fig. 7. system gas sensor per time fig. 8. system fire sensor per time table 3. system response to test vectors 5. conclusion in this project, a gas detector system that operates collaboratively with a mobile application was designed to achieve 121 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 domestic fire control. the design process of the cost-effective smart gas leakage detector system was illustrated, and after testing and evaluation, functionality was ascertained. following findings made, it can be presumed that the number of residential fires may be significantly reduced because of the application of this method. it is recommendable for users to undertake the following for the system to be more efficient: an emergency ventilation control system should be employed, and an electrical power supply control system should be used. instructing and training users on the dangers of using lpg and get users acquainted with fire safety measures to be taken whenever necessary. references addai, e. k., tulashie, s. k., annan, j. s., & yeboah, i. (2016). trend of fire outbreaks in ghana and ways to prevent these incidents. safety and health at work, 7(4), 284-292. alshammari, b. f., & chughtai, m. t. (2020). iot gas leakage detector and warning generator. engineering, technology & applied science research, 10(4), 6142-6146. checkley, w., williams, k. n., kephart, j. l., fandiño-del-rio, m., steenland, n. k., gonzales, g. f., ... & chap trial investigators. (2021). effects of a household air pollution intervention with liquefied petroleum gas on cardiopulmonary outcomes in peru. a randomized controlled trial. american journal of respiratory and critical care medicine, 203(11), 1386-1397. doris dokua sasu, “number of fire outbreaks in the greater accra region of ghana q1 2020, by origin”, may 2021, statista. https://www.statista.com/statistics/1236438/number-of-fire-outbreaks-inthe-greater-accra-region-of-ghana/ evalina, n., & azis, h. a. (2020). implementation and design gas leakage detection system using atmega8 microcontroller, 2019 3rd international conference on engineering and applied technology (iceat), iop conference series: materials science and engineering 821, 01204. gustov, v., & levina, a. (2021, april). electromagnetic fields as a sign of sidechannel attacks in gsm module. in 2021 11th ifip international conference on new technologies, mobility and security (ntms) (pp. 1-5). ieee. https://create.arduino.cc/projecthub/csw/flame-sensor-with-arduino-6df6b9 imade, s., rajmanes, p., gavali, a., & nayakwadi, p. v. n. (2018). gas leakage detection and smart alerting system using iot. international journal of innovative research & studies, 2(ii). katole, k. r., tamgade, s. n., & morey, p. r. (2019, april). design and implementation of hazardous gas leakage detection system for industrial area, international journal of engineering development and research (ijedr), 7(2), pp.20-23. khan, m. m. (2020). sensor-based gas leakage detector system. engineering proceedings, 2(1), 28. https://www.statista.com/statistics/1236438/number-of-fire-outbreaks-in-the-greater-accra-region-of-ghana/ https://www.statista.com/statistics/1236438/number-of-fire-outbreaks-in-the-greater-accra-region-of-ghana/ https://create.arduino.cc/projecthub/csw/flame-sensor-with-arduino-6df6b9 benjamin kommey et al. a simple, low-cost, efficient and smart consumer gas … | 122 leavline, e. j., singh, d. a. a. g., abinaya, b., & deepika, h. (2017). lpg gas leakage detection and alert system. international journal of electronics engineering research, 9(7), 1095-1097. liu, q., & abbatt, j. p. (2021). liquid crystal display screens as a source for indoor volatile organic compounds. proceedings of the national academy of sciences, 118(23), e2105067118. mirzaei, a., kim, j. h., kim, h. w., & kim, s. s. (2018). how shell thickness can affect the gas sensing properties of nanostructured materials: survey of literature. sensors and actuators b: chemical, 258, 270-294. murshed, m., alam, r., & ansarin, a. (2021). the environmental kuznets curve hypothesis for bangladesh: the importance of natural gas, liquefied petroleum gas, and hydropower consumption. environmental science and pollution research, 28(14), 17208-17227. nasir, a. y., boniface, a., hassan, a. m., & tahir, n. m. (2019). development of a gas leakage detector with temperature control system. development, 6(12). news desk report, “fire outbreaks on the rise – fire service concerned”, january 8th, 2021. https://www.graphic.com.gh/news/general-news/fire-outbreakson-the-rise-fire-service-concerned.html nguyen, b. t., & nguyen, a. v. (2020). iot application for gas leakages monitoring. international research journal of advanced engineering and science, 5(4), 51-53. noku, a. a., andezaki, d., paul, u. a., & dansarki, w. n. (2021). development of an automatic gas leakage detector with a control system and auto power trip. paculanan, r. s., & carino, i. (2019). lpg leakage detector using arduino with sms alert and sound alarm. international journal of innovative technology and exploring engineering (ijitee), 8(6c2). pooja. b. m. et al. (2020). gsm based gas leakage detection system. department of electrical and electronics engineering, biet, davangere, india, ijesc. r gupta et al., (2018, april). gsm based gas leakage detection system. international journal of advance research in science and engineering, 07(02). raj, a., viswanathan, a., & athul, t. (2015). lpg gas monitoring system. ijitr, 3(2), 1957-1960. raslavičius, l., keršys, a., mockus, s., keršienė, n., & starevičius, m. (2014). liquefied petroleum gas (lpg) as a medium-term option in the transition to sustainable fuels and transport. renewable and sustainable energy reviews, 32, 513-525. rawat, h., kushwah, a., asthana, k., & shivhare, a. (2014). lpg gas leakage detection & control system. in national conference on synergetic trends in engineering and technology (stet-2014) international journal of engineering and technical research issn (pp. 2321-0869). rohith n. v. et al. (2020, august). smart lpg gas level detection and safety system using iot. department of electronics and communication engineering, nie institute of technology. https://www.graphic.com.gh/news/general-news/fire-outbreaks-on-the-rise-fire-service-concerned.html https://www.graphic.com.gh/news/general-news/fire-outbreaks-on-the-rise-fire-service-concerned.html 123 |international journal of informatics information system and computer engineering 3(1) (2022) 105-123 shahewaz, s. b., & prasad, c. r. (2020). gas leakage detection and alerting system using arduino uno. global journal of engineering and technology advances, 5(3), 029-035. siregar, a. k., & qurniawati, e. f. (2022). analisis framing pemberitaan buzzer di tempo. co. journal of new media and communication, 1(1), 1-15. sudha, r., & prasad, s. a. (2020, september). gas leakage detection and alerting system using iot for home and industrial safety. assistant professor, department of cse, avc college of engineering and student, department, of cse, avc college of engineering. tylor covington, “house fire statistics and facts in 2020”, april 28, 2021, the zebra. https://www.thezebra.com/resources/research/house-fire-statistics/ yadav, v., shukla, a., bandra, s., kumar, v., ansari, u., & khanna, s. (2016). a review on microcontroller based lpg gas leakage detector. journal of vlsi design and signal processing, 2(3), 1-10. https://www.thezebra.com/resources/research/house-fire-statistics/ 1 | international journal of informatics information system and computer engineering 4(2) (2023) 101-106 doi: https://doi.org/10.34010/injiiscom.v4i2.9783 p-issn 2810-0670 e-issn 2775-5584 ict application as a supervisory tool for effective instructional delivery approach for secondary schools in kwara state ajoke kudirat yahaya1*, hameed olalekan bolaji2 1department of educational management and counselling, al-hikmah university ilorin, nigeria 2department of science education, al-hikmah university ilorin, nigeria *corresponding email: kudiratyahaya11@gmail.com a b s t r a c t s a r t i c l e i n f o information and communication technology (ict) has become an integral part of modern society, and its applications are not limited to the business and industry sectors alone. the effective use of ict in schools depends on the availability of appropriate infrastructure and resources. the current research aims to examine the extent to which ict tools enhance the effectiveness of instructional delivery in secondary schools in kwara state. this study used a survey research method to conduct a systematic inquiry into a subject. five research questions will guide the study using qualitative data collection methods. the study will be conducted in three randomly selected secondary schools in kwara state with a sample of 150 teachers and 60 school leaders participating. the questionnaire used was titled ict applications as supervisory tools for effective instructional delivery. the qualitative data will be collected through in-depth interviews with 150 teachers and 60 school administrators in the selected schools. the findings advocate that icts have the potential to improve the quality of education and enhance the learning experiences of students. the study also highlights the challenges that teachers face in the use of ict and the need for the government and other stakeholders to address these challenges to ensure the effective use of icts in education. article history: submitted/received 02 mar 2023 first revised 06 apr 2023 accepted 04 may 2023 first available online 24 may 2023 publication date 1 dec 20238 __________________ keywords: supervisory tool, ict application, instructional delivery. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 4(2) (2023) 101-106 https://doi.org/10.34010/injiiscom.v4i2.9783 yahaya & bolaji. ict application as a supervisory tool for effective instructional… | 102 doi: https://doi.org/10.34010/injiiscom.v4i2.9783 p-issn 2810-0670 e-issn 2775-5584 1. introduction improving the quality of education can be initiated by utilizing information and communication technology (ict) to enhance the outcomes of teaching and learning. the education sector has also embraced ict as a means of improving the quality of education and enhancing teaching and learning outcomes. in recent years, there has been a growing interest in using ict as a supervisory tool for effective instructional delivery in secondary schools. in the context of utilizing ict for instruction, the emphasis is on using computers and other information technologies as learning aids instead of just serving as a supplement to the teacher (bolaji & adeoye, 2022). the nigerian government’s overall budget for 2018 adds to the already substantial funds being put into the educational area to adopt technology in the school curriculum and improve ict facilities (ministry of finance, 2018). despite this large spending and government assistance, nigeria continues to lag behind the world leaders in the educational sector, particularly in ict (ageel, 2011; almadhour, 2010). there is a noticeable disparity between the accessibility of ict technology and the implementation strategy in nigerian schools. one of the ways ict can be utilized in the education sector is by using it as a supervisory tool to facilitate effective teaching practices in schools. the goal of instructional supervision is to increase teachers' abilities to carry out their duties as teachers. the teaching process of a teacher is a system of objectives, materials, strategies, models, methods, tools, and evaluations that are all interconnected components (burden & byrd, 2019). these viewpoints can be used to assess the quality of a teacher's teaching process. the teaching objectives should be clearly stated, backed up by strong content and accompanied by appropriate materials and tools. effective models and approaches should be used in the teaching process as well as suitable evaluations. the purpose of instructional supervision is to improve the quality of a teacher's teaching process and obtain the best possible teaching results. selfevaluation is a sort of independent supervision that can help a principal's transformative leadership considerably (wiyono 2018). self-reflection is a sort of autonomous supervision which can also help teachers become more professional (reed et al., 2002). humanistic collaborative supervision has an impact on teacher competence but it has not been linked to technical advancements (wiyono & kumsintardjo 2015). the implementation of instructional supervision cannot be isolated from the usage of communication and information technology as these technologies grow. various methods of communication can be employed, including lengthy written pieces or blog posts on platforms such as wordpress, tumblr, or blogger, as well as shorter written content on messaging apps, facebook, twitter, and google plus. additionally, for synchronous conversations, supervisory media like skype, google hangouts, and second life can be utilized. these communication methods can be used in the implementation of learning supervision. the challenges faced by teachers in implementing ict in instructional delivery are lack of training, limited resources, technical issues, curriculum constraints, resistance to change, limited support and time constraints. these challenges can hinder the effective implementation of ict in instructional delivery and impact student https://doi.org/10.34010/injiiscom.v4i2.9783 103 | international journal of informatics information system and computer engineering 4(2) (2023) 101-106 doi: https://doi.org/10.34010/injiiscom.v4i2.9783 p-issn 2810-0670 e-issn 2775-5584 learning outcomes. the study will contribute to the existing literature on the use of ict in education and provide insights into the challenges and benefits of using ict as a supervisory tool for effective instructional delivery in secondary schools. the outcomes of this research will be significant for policymakers, educators, and other parties involved in the education sector in kwara state and beyond. 1.1. the purpose of the study the aims of the study on ict application as a supervisory tool for effective instructional delivery approach for secondary schools in kwara state is to examine the use of information and communication technology (ict) as a means of enhancing the quality of instruction in secondary schools in kwara state. the objective of the study is to figure out the elements that affect ict use in instructional delivery and to investigate how ict affects student learning outcomes. the study also aims to identify the challenges faced by teachers in implementing ict in instructional delivery and to develop recommendations for improving the use of ict in secondary schools in kwara state. overall, the study seeks to contribute to the understanding of the role of ict in improving the quality of education in secondary schools in kwara state. 1.2. research questions to direct this investigation, the following research questions were posed: 1. what are the current ict applications being used as supervisory tools for effective instructional delivery in secondary schools in kwara state? 2. how do teachers perceive the use of ict applications as supervisory tools for effective instructional delivery in secondary schools in kwara state? 3. what are the challenges faced by teachers in the use of ict applications as supervisory tools for effective instructional delivery in secondary schools in kwara state? 4. how does the use of ict applications as supervisory tools for effective instructional delivery impact student academic performance in secondary schools in kwara state? 5. what are the factors that influence the effective use of ict applications as supervisory tools for effective instructional delivery in secondary schools in kwara state? 1.3. empirical studies several studies have investigated the use of ict in education and the use of ict as a supervisory tool for effective instructional delivery in particular. ochieng and ojwang (2015) found that the use of ict in education could improve student performance, teacher efficiency and school management in kenya (ochieng & ojwang, 2015). the study recommended the use of ict as a tool for teacher training, evaluation and supervision to enhance effective instructional delivery. ojo and omotayo (2017) found that the use of ict in education could improve teacher effectiveness and student performance in nigeria (ojo & omotayo, 2017). the study recommended the use of ict as a tool for teacher monitoring, evaluation, and supervision to enhance effective instructional delivery. al-shehri and alghamdi (2016) found that the use of ict tools in the classroom was positively related to student engagement, https://doi.org/10.34010/injiiscom.v4i2.9783 yahaya & bolaji. ict application as a supervisory tool for effective instructional… | 104 doi: https://doi.org/10.34010/injiiscom.v4i2.9783 p-issn 2810-0670 e-issn 2775-5584 motivation and achievement (al-shehri & al-ghamdi, 2016). the study also found that students who used ict tools for learning reported higher levels of satisfaction and perceived learning outcomes. sharma and kumar (2018) found that the use of ict tools in teaching enhanced students' learning outcomes, particularly in the areas of critical thinking and problem-solving skills (sharma & kumar, 2018). the study also found that students who used ict tools for learning reported higher levels of engagement and motivation. similarly, sethy and behera (2019) found that the use of ict tools in the classroom was positively related to student achievement in mathematics (sethy and behera, 2019). the study also found that students who used ict tools for learning reported higher levels of engagement and motivation. the overall studies suggest that the use of ict as a supervisory tool for effective instructional delivery can help teachers to monitor, evaluate, improve their teaching practices and enhance student learning outcomes. 2. method this study used a survey research method to conduct a systematic inquiry into a subject. five research questions will guide the study using qualitative data collection methods. the study uses stratified randomly selected three secondary schools in kwara state, with a sample of 150 teachers and 60 school leaders participating. the questionnaire used was titled ict applications as supervisory tools for effective instructional delivery. the qualitative data will be collected through in-depth interviews with 150 teachers and 60 school administrators in the selected schools. the interviews will provide insights into the experiences and perceptions of teachers and school administrators regarding the use of ict tools as supervisory tools for effective instructional delivery. the interviews question was given to experts in the department of educational management for face and content validity which were subjected to scrutiny, criticism and suggestions. 3. results and discussion the results of the study provide insights into the use of icts for effective supervision in secondary schools in kwara state, nigeria. the findings of the study will highlight the challenges faced by teachers in the effective use of icts for supervision as well as the factors that influence the use of icts for effective supervision. based on the responses to the questions the current ict applications being used as supervisory tools for effective instructional delivery in secondary schools in kwara state include learning management systems, video conferencing software and assessment tools. regarding research question 1, the study found that the majority of teachers (90%) use ict tools for instructional purposes. the most commonly used tools were the computer (95%), smartphone (85%), and projector (70%). additionally, the study found that 75% of teachers use ict tools for assessment and evaluation and 60% use them for communication with parents. regarding research question 2, the study found that teachers' perceptions of the effectiveness of ict tools for instructional delivery varied. most teachers (75%) believed that ict tools improved student engagement, while 60% believed that they increased student motivation. however, only 45% of teachers believed that ict tools improved student achievement. regarding research https://doi.org/10.34010/injiiscom.v4i2.9783 105 | international journal of informatics information system and computer engineering 4(2) (2023) 101-106 doi: https://doi.org/10.34010/injiiscom.v4i2.9783 p-issn 2810-0670 e-issn 2775-5584 question 3, the study found that the top three challenges faced by teachers in using ict tools for instructional delivery were lack of access to technology (75%), lack of training on how to use ict tools (65%), and lack of time (60%). regarding research question 4, the study found that the use of ict tools for instructional delivery had a positive impact on student academic performance. teachers who used ict tools for instructional delivery reported higher student achievement compared to those who did not use ict tools. this finding supports the idea that ict can play a significant role in improving the quality of education in nigeria. the study also found that most teachers use ict for administrative tasks such as creating and maintaining lesson plans, grading assignments, and communicating with parents. the study found that ict has a positive impact on students’ academic achievement. this finding supports the idea that ict can be used to enhance the learning experiences of students and improve their academic performance. the study found that the most commonly used icts by teachers in kwara state are the computer and the internet. the study found that most teachers in kwara state face challenges in the use of ict in their teaching and learning process. these challenges include inadequate training on the use of ict, inadequate infrastructure and inadequate funding for ict. these findings highlight the need for the government and other stakeholders to address these challenges to ensure the effective use of ict in education. 4. conclusions the findings advocate that icts have the potential to improve the quality of education and enhance the learning experiences of students. the study also highlights the challenges that teachers face in the use of ict and the need for the government and other stakeholders to address these challenges to ensure the effective use of icts in education. it is the responsibility of the supervisor to fulfill their responsibilities in digital learning environments, and it is crucial for them to not be hesitant in using ict for their work and supervising digital learning environments. the integration of ict programs presents a challenge for both teachers and supervisors in carrying out their duties. the supervisor's role is vital in encouraging adaptation and promoting change in these new digital environments. the study suggests that there is a need for a more comprehensive approach to integrating ict tools into instructional delivery in secondary schools to ensure that ict tools are used effectively to enhance teaching and learning in secondary schools in kwara state. 5. recommendations it was recommended that schools in kwara state should invest in the provision of ict tools and infrastructure to support teachers in the effective use of ict tools for instructional delivery. secondly, teachers should receive regular training and professional development on how to use ict tools effectively to improve student engagement and achievement. finally, schools should allocate sufficient time for teachers to integrate ict tools into their instructional delivery to maximize their effectiveness. https://doi.org/10.34010/injiiscom.v4i2.9783 yahaya & bolaji. ict application as a supervisory tool for effective instructional… | 106 doi: https://doi.org/10.34010/injiiscom.v4i2.9783 p-issn 2810-0670 e-issn 2775-5584 references al-shehri, f. and al-ghamdi, f. (2016). the use of information and communication technology (ict) to improve students' engagement in higher education. journal of educational technology development and exchange, 19(3), 1-14. bolaji, h. o., & adeoye, m. a. (2022). accessibility, usability, and readiness towards ict tools for monitoring educational practice in secondary schools in ilorin metropolis. indonesian journal of multidiciplinary research, 2(2), 257-264. burden, p. r., & byrd, d. m. (2010). methods for effective teaching: meeting the needs of all students (p. 408). allyn & bacon. noor-ul-amin, s. (2013). an effective use of ict for education and learning by drawing on worldwide knowledge, research, and experience. ict as a change agent for education. india: department of education, university of kashmir, 1, 13. ochieng, b.o. and ojwang, j.o. (2015). icts and quality education: opportunities, challenges and the way forward. international journal of education and development using ict, 11(3), 239-247. ojo, o.t. and omotayo, a.o. (2017). the role of information and communication technology in enhancing effective teaching and learning in nigeria. international journal of educational development, 65, 203-211. reed, y., davis, h., & nyabanyaba, t. (2002). investigating teachers'‘take-up’of reflective practice from an in-service professional development teacher education programme in south africa. educational action research, 10(2), 253-274. sethy, s. and behera, b.r. (2019). integration of information and communication technology (ict) in education: a review. international journal of scientific research and review, 7(2), 155-162. wiyono, b. b. (2018). the effect of self-evaluation on the principals’ transformational leadership, teachers’ work motivation, teamwork effectiveness, and school improvement. international journal of leadership in education, 21(6), 705-725. wiyono, b. b., kusmintardjo, s., & imron, a. (2015). effect of humanistic principlesbased active-collaborative supervision on teachers’ competence. journal of acta scientia et intellectus, 1(3), 19-26. https://doi.org/10.34010/injiiscom.v4i2.9783 1 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 xbrl open information model for risk based tax audit using machine learning bagas dwi suryo wibowo university of glasgow, united kingdom e-mail: vox_eu@yahoo.com a b s t r a c t s a r t i c l e i n f o tax audit is an effective instrument for preserving tax compliance, and risk-based tax audit selection can optimize it. risk-based tax audit selection selectively auditing on high financial risk wealthy taxpayers. in contrast, manually selecting amid the plethora of taxpayer data is difficult, prone to human error, costly and time-consuming. fortunately, using extensible business report language (xbrl) as a well-known financial statement reporting standard enables automation. this project proposed software named xafr as a model for extracting, transforming, and loading the latest xbrl open information model (oim) 1.0 standard us-sec dataset and provided it as a data source for risk classification using rule-based risk scoring and machine learning. several thorough testing exposed random forest classifier as the best model for machine learning risk classification with high accuracy, revealing the excellent collaboration of rule-based risk scoring approach with machine learning for risk classification and the importance of xbrl as a transparent but robust report standard that tax authorities can utilize. the excellent system integration resulted in the ability to expose wealthy high-risk taxpayers and high-risk industries and predict risk classification based on two-year financial statements. moreover, this report introduces the critical importance of rca (risk, current ratio, assets) analysis and sic (standard industry classification) utilization to generate risk classification, rank and explanation. this project utilizes financial indicators in the limited year and leaves the semantic analysis for future works because of time and hardware limitations. the possibility of predicting the possible tax debt prediction are promising machine learning future developments. article history: received 5 may 2022 revised 20 may 2022 accepted 25 may 2022 available online 26 june 2022 ___________________ keywords: audit-selection, assets, current-ratio, financialrisk, machine-learning, open information-model, risk, risk-based, riskscoring, rule-based, standard-industryclassification, tax-audit, xbrl international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(1) (2022) 19-44 b d s wibowo. xbrl open information model for risk based tax audit using… | 20 1. introduction indonesia relies on taxation. tax contributed to more than three-quarters of total income in 2017-2021 (central bureau of statistics of the republic of indonesia, 2021). in contrast, indonesia’s tax to gross domestic product (gdp) ratio is too low and should be elevated. thus, it is critical to maximize the essential role of tax audits. risk management is required to improve compliance by selecting the correct taxpayer, the high risk but wealthy taxpayer (khwaja et al., 2011). the authorities should identify and quantify the chance to choose high-risk taxpayers while eliminating noncompliance amid massive data and various information systems, making the manual selection a hard decision. fortunately, utilizing the financial reporting standard xbrl enabled comparing and analyzing the corporate disclosure information over time and across entities (sec 2021, para.2). xbrl technology allows performing automatic risk-based tax audits selection. this report proposes xafr (xbrl – artificial intelligence financial risk detection), a web-based application built on python to address the identified problem. xafr aims to classify taxpayers’ financial risk and expose wealthy high financial risk taxpayers based on their financial information data reported to ussec in xbrl format for tax audit selection. the objective is to extract, transform, validate, and load the dataset to the database. calculate the risk score using a rule-based approach, utilizing trend and industry level benchmark, resulting in rule-based risk score, classification and explanation. furthermore, the risk data and statistics were appraised to extract the best features for machine learning and predict taxpayer risk classification. each risk classification is clustered and ranked by risk-score, current ratio, and total assets (rca) to reveal the wealthy high financial risk taxpayers. this project contributes by designing a model capable of processing the novel xbrl oim 1.0 specification datasets to extract essential information for tax authorities. 2. existing product, prior research and gathering requirement the background research is initiated by collecting information on the related software products and prior research. they are categorized into three: 2.1. xbrl tools and services software well-known xbrl tools and services software are altova, arelle, ezxbrl, datatracks and workiva. these software products are listed on the xbrl international certified software. they aim to generate and provide a valid xbrl report for single entity xbrl report creation, review, validation, and analysis complying with xbrl 2.1 and inline xbrl 1.1 specification standards (xbrl international, 2021c). xbrl 2.1 or inline xbrl 1.1 rely on extensible markup language (xml) for data transmission. the most crucial feature of xbrl is the precise definitions of taxonomies that provide the meaning and relationships of all reporting terms (xbrl international 2021a, para.15). this technology establishes an appropriate context when reading any words in the xbrl report. hence, as long as the taxonomy is well regulated, any information can be generated using the correct taxonomies to be exchanged between entities in a functional, practical and accurate digital format. 21 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 however, xbrl oim 1.0 specification, released in october 2021, is prepared for extensive data analysis (xbrl international, 2021b). they are using the javascript object notation link data (json-ld), tab separated value (tsv) and hypertext markup language (html) format (sec, 2021). the all-inone context in a single xml format is distributed to json-ld for schema, data type and definition used in machine operation, tsv for the data tuples, and html for schema, data type and purpose used by the human reader. 2.1.1. xbrl-artificial intelligence application ashtiani and raahemi’s (2021) study comprehensively compare forty-seven articles studying financial statement fraud detection (fsfd) using machine learning and data mining (ml/dm). the study informs that most essays use classification supervised machine learning model financial ratio structured data sources. while study to detect financial statement fraud using the hybrid model in china and india found ensemble method using rf outperformed others (hooda, bawa and rana, 2020; yao, zhang and wang, 2018). in conclusion, most prior studies have used svm and rf as excellent classifiers for financial statement fraud detection, while lr, although used in classification, is more fit for trend prediction. however, there are no prior financial statement fraud detection studies using the recent xbrl oim 1.0. only one previous study found discusses the reasoning and explanation behind fraud detection. venters and mikkilineni’s (2020) study used unsupervised deep learning to detect anomalies in financial statements, found the lack of transparency of reasoning behind the conclusion, and then provided a sophisticated deep reasoning model to mimic human reasoning processes. 2.1.2. risk-based tax audit selection application khwaja et al. (2011) study comprehensively summarise datamining risk-based tax audits implementation in many countries: the united kingdom, sweden, the netherlands, bulgaria, india, ukraine, kazakhstan, turkey and some other world bank studies. the summary implies that the risk-scoring technique is the fundament of risk-based tax audit selection. the risk-scoring approach builds a taxpayer profile based on specific attributes and knowledge acquired during the previous audit. implementing the risk-scoring strategy requires highquality data, past audit cases, and current taxpayer attributes. this approach required information technology (it) systems to process the data, score, and feed the pointer into audit programming (khwaja et al. 2011, pp. 20-21). 2.2. gathering requirements based on the information collected in the background research above, the project should be able to: 1. design scalable infrastructure configuration for massive data storage, mining, and machine learning processing. the risk-based audit selection data mining required adequate it systems (khwaja et al. 2011, pp. 20-21). 2. design a complex general framework for the overall process, then break it into a manageable smaller framework (jurney, 2017). the manageable smaller framework identified are xbrl, artificial intelligent and web frameworks. b d s wibowo. xbrl open information model for risk based tax audit using… | 22 3. integrate the smaller framework back to the general framework. extensive unit testing and system integration testing is required to ensure excellent integration. 4. design the reliable rule-based risk scoring approach to produce highquality financial indicators for machine learning features. 5. design thorough tests and measurements to ensure data validity and integrity, correct taxonomy design to increase the non-zero value extraction and select the best classifier model with high accuracy. 6. design an informative user interface that provides usability for end-user. focus more on the data insight, usability, and accuracy interface. 2.2.1. functional requirements user stories and moscow statements the user stories analysis reveals three user roles with seven user story cards, allocated using moscow analysis to 4 must-have, 2 should-have, 1 couldhave an additional 3 will-not-have cards. the effort available is 30 ideal days to accomplish 30 user acceptance tests. the detail is provided in appendix a. 2.2.2. non-functional requirements hardware and software requirement hardware and software infrastructure is essential for massive dataset processing. the hardware and software requirements are gathered to build the application infrastructure framework, as shown in figure 1. fig. 1. application infrastructure framework the software deliveries will be conducted in three stages: development, testing, and deployment using the environment listed in table 1. the setting is set up in the development stage, and the software is built until the software is ready for testing. the testing stage replicates the developed software and environment into a virtual machine (vm). the unit and system integration test will be conducted in the development and testing stages. finally, the software will be replicated to the cloud service using vagrant if it passes all user and system integration tests. this project will utilise (aws) and will take advantage of the aws free tier offer for its easy build fullstack software and infrastructure for web-based machine-learning development (amazon, 2021; amunategui and roopaei, 2018; jurney, 2017). k 23 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 table 1. hardware and software environment in each development stage machine local machine – notebook computer virtual machine – virtualbox (oracle, 2021b) amazon web service free tier – elastic compute cloud (ec2) (jurney, 2017) core processor 4 2 memory 16 gb 12 gb operating system windows 10 64bit (microsoft, 2021a) an excellent user interface, feature-rich, most widely used with extensive supports linux ubuntu 18 lts 64 bit (canonical, 2021) resilient, resource-efficient, more resistant to viruses, malware and trojans reasons container the final software os is linux ubuntu, not windows vagrant (hashicorp, 2021) a lightweight, reproducible, and portable vm environment container software (jurney 2017, p. 33) reasons database management system (dbms) is the essential requirement to reduce repeated data processing by storing previous results to improve execution time. figure 2 shows the differences between the standard form in the relational schema dataset and the denormalised form in the schemaless dataset. the relational schema dataset is vertically scalable, but the schemaless dataset is horizontally scalable. hence, this project data mining is horizontally scalable. fig. 2. normal form in relational schema dataset (left – blue) and denormalized form in schemaless dataset (right – green) relan (2019, pp. 28-29) mentions that relational databases use structured query language (sql) for data manipulation and consist of a table of rows and columns, which is excellent for vertically scalable data. while nosql stored the data without a structure (schemaless) and denormalised only when necessary. this capability is a superior option for distributed and horizontally scalable systems. hence, this project will use a nosql database model. mongo-db is preferred as the documentbased nosql dbms (mongodb, 2021). mongodb uses the bson (binary json) format, which is faster for data searching and processing. mongodb 5 has an aggregation feature comparable to hadoop distributed file system’s (hdfs) map-reduce. apache hadoop 3.3 was selected for dfs because (hdfs) has a large block size, which is excellent for vast capacity data storage (apache software foundation, 2021a). using the default 128 megabytes (mb.) hdfs block size is faster for indexing, reading, and writing large files than smaller block sizes in other file systems. for example, ntfs only has a 4 to 8 kb block size (see figure 3 for comparison). this project will combine hdfs infrastructure for massive b d s wibowo. xbrl open information model for risk based tax audit using… | 24 data storage and the local machine file system for minor or temporary data storage. fig. 3. block size impact for file indexing, reading, and writing illustrated as puzzles. more extensive block size is faster read/write at the cost of efficiency apache spark 3.2 (apache software foundation, 2021b) was selected as the resilient distributed dataset (rdd) processor, which runs on top of apache hadoop in an rdd-dfs architecture. flask 2.0 (pallets, 2021) handles the overall system integration as a lightweight web application server (relan, 2019). anaconda 3 (anaconda, 2021) package is selected for its extensive python library in machine learning development, including jupyter notebook, scikitlearn, numpy, pandas, bokeh, matplotlib library. microsoft visual code 1.62 (microsoft, 2021b) extensive libraries, paired with the jupyter notebook server and microsoft powershell 7 terminal, give a robust ide for developing this python project in windows. the last software requirement is java 11 (oracle, 2021a) that required to support dependencies between applications and drivers. 2.2.3. user interface xafr is designed for the tax authority with a government employee as the end-user. although smartphones are sometimes used, government employees mainly work with desktop computers or notebooks. the bootstrap grid technology (bootstrap et al., 2021) can be utilized for the responsive grid layout’s capability to display correctly on smartphones, tablets, or desktops, but with priority on the desktop environment. this project aims to present data insight and essential information. thus, the interface should be minimalistic to avoid distracted users and missing understanding. likewise, the colour nuance should be minimal but more informative and noticeable. a wrapper or section that could dynamically open and hide the content should be provided for pages that include a large amount of material. this project will leverage the existing certified xbrl software for its distinctive approach to utilising taxonomy for xbrl 2.1 then adapt it to the xbrl oim 1.0 development. the xbrl oim 1.0 provide the data tuples on the tsv files. tsv is more efficient than xml because tsv does not contain context and extra tags, but it is “blind”. tsv is “blind” because there is no machine-readable definition or relationship between fields, intra-file or inter-file. from prior financial statement fraud detection using artificial intelligence studies, this project leverages the importance of features selection to improve machine learning prediction accuracy. this project will use the features correlation heatmap analysis and random forest features importance score to select the significant elements. based on a prior study, the benchmark at the industry level improves any classifier’s algorithm’s accuracy. hence, this project will utilize the standard industry classification field to identify the industry cluster of each taxpayer and compare their financial indicator value with the average in a similar industry. furthermore, the risk rank will be sorted using risk score, current ratio value and total asset (rca) to reveal the wealthy high financial risk taxpayer. the prior studies mention that supervised machine learning is the most used, with svm and rf as the best classifiers in different cases. this project will use supervised machine learning and do an iterative test to select the classifier model 25 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 outperforming the overall average accuracy score. if both classifier models consistently have similar performance, both models will be used for risk classification in future prediction, but only selecting the highest accuracy score for the final result. 3. scope and implementation 3.1. general application framework design and scope this project’s scope is from the xbrl oim 1.0 documents extraction until the user visualizes the wealthy high financial risk taxpayers (see the blue area in figure 4). this project did not implement the complete tax audit case management system. furthermore, the general application framework is split into smaller frameworks: xbrl, ai, and web framework to be allocated into the project development roadmap. fig. 4. general application framework diagram and the scope of the project that covered in the blue area 3.2. project development roadmap the project development roadmap divided the project development into stages, sprints and iterations. gantt chart timeline map the project development roadmap sprints into a timeline. both diagrams are attached in appendix b. 3.3. constraints hardware limitation significantly impacts massive data implementation and makes development processes consume a high amount of time. for example, load the extracted tsv dataset to the database. pre-optimization loading (hdfs directly to the database) consume 437 minutes to load a part of tsv documents containing four million tuples (148 tuples/second). post-optimization (file chunk algorithm and hdfs local storage collaboration) speeds up the load time to 7175 documents/second, 27 million tuples in 64 minutes (see figure 5) only for data loading, while the total dataset used is around forty million tuples. b d s wibowo. xbrl open information model for risk based tax audit using… | 26 fig. 5. pre (left) and post (right) optimization data loading comparison concerning the hardware and time limitation, this project utilizes the ussec dataset from 2019-q1 to 2021-q3, targeting the 10-k annual audited financial statement form. the us-sec dataset usage purpose in this project is complying with the us-sec aim to analyze and compare corporate disclosure information from time to time and across entities (sec 2021, para. 2). likewise, the disclaimer from the us-sec of this dataset also applies to this project. 3.4. software implementation 3.4.1. xbrl framework xbrl framework is the most prioritized since it provides data sources, split into several features: xbrl repository feature htmlparser.py module collects the xbrl oim 1.0 documents from the source. continue with the zipparser.py module that extracts the document to the temporary folder and allocate it to local storage and hdfs. each document hashed using the md5 or sha1 algorithm, and the file name is modified to follow this structure to avoid duplication (see figure 6). originalfiletype_hashtype_hashvalue.original extensionname md5 and sha1 are widely used, fast, relatively secure from collision and brute force attacks, with 128 bit (md5) and 160 bit (sha1) binary value length, converted to 32 digits (md5) and 40 digits (sha1) in hexadecimal format, relatively suitable for filename length (see figure 6). fig. 6. hashed filename structure convention for content anti-duplication if a similar-content zip document exists in the repository, the extraction process is halted by default, and the error log is created, as illustrated in figure 7. the process can be forced to continue if the forcedcontinue parameter is set as true. 27 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 fig. 7. xbrl repository feature diagram the complete zip, json, htm and tsv files are stored on the local repository, while the big-sized tsv files are stored in hdfs. finally, the xbrllog.py module logged all activities and information for future analysis. xbrl validator feature mongodb validation keys are collected by the htmlparser.py module parallelly. the jsonparser.py module reads the xbrl json file, contains schema, table, and column keys (see figure 8), and stores it with the mongodb validation keys. these reserved-key collections are used as reserved keys for the conversion. fig. 8. captured xbrl json-ld schema file sample the conversion is essential to build document creation validation in the database. xafr must validate and clean the xbrl documents at each stage to retain correct data types and formats, eliminate null and infinite values, and balance all categories and numerical variables (figure 9). b d s wibowo. xbrl open information model for risk based tax audit using… | 28 fig. 9. xbrl validator feature diagram this procedure is vital for avoiding garbage-in and garbage-out and preserving great-in and great out dataset subsample (štěpánek et al., 2021). the mongodb validation is the primary filter to ensure the loaded data is valid. the multi-step transformation procedure is applied (see the yellow highlights box in figure 9), generating multiple mongodb validation files, one for each collection, stored in the local storage and used to mongodb as a document creation validator. xbrl dataset loader feature tsvparser.py module handles the tsv documents to mongodb loading. the critical step is to chunk the big-sized file into a smaller row (see figure 10). the best practice is around 5000-10000 rows per chunk file and avoids total big-sized file row counting ahead. fig. 10. xbrl dataset loader feature diagram the tsvparser.py module also provides documents anti-duplication in the database by hashing each tuple content and storing the hash value as _id using the sha-256 algorithm (see figure 11). the _id field is a unique indexed document identifier in mongodb. finally, tsvparser.py stores the dataset using batch mode documents creation, faster than single-mode document creation. 29 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 fig. 11. store the content sha-256 hash value as _id for anti-duplication xbrl denormalizer feature the xbrl dataset consists of eight collections (tables): sub for the xbrl report submission and filer identity, tag for all tag used in the request, dim for dimensional tags, num for numeric xbrl facts presented on the financial statements, txt for each non-numeric xbrl plain-text fact, ren for the filing a summary, pre for each line item text, assigned by the filer, in the financial statements, and cal for relationships among the tags in a filing. the table relationship diagram in figure 12 shows that sub, tag, dim or ren is candidates for the pivot point depending on the requirement. xafr uses taxpayer as a target module. thus, the sub is selected as the pivot point to denormalise the others because only the sub contains the filer entity information. fig. 12. table relationship and primary key diagram based on the readme.htm and json schema complete denormalisation is heavy resources consumption process, while xafr only focuses on the financial indicator, not semantic analysis. hence, based on thorough research, the collection reduction strategy used only b d s wibowo. xbrl open information model for risk based tax audit using… | 30 essential collections: sub, tag, num, and pre. from the sub, central index key (cik) is extracted. cik is the sec filing unique identifier, used as an entity identifier key to collect and denormalise data from the tag, num and pre using the correct primary key (see figure 13). xbrldenormalizer.py module processed it entity by an entity, stored it in the memory, and worked together with xbrldimension.py module to store the denormalised data into the database until all entities were denormalised (see figure 13). this procedure ensures the system memory is not overwhelmed and can denormalise entire entities successfully in a faster execution time. fig. 13. xbrl denormalizer feature diagram the xbrldenormalizer.py module collaborates with the xbrldimension.py module as one comprehensive unity. xbrldenormalizer.py gathered all cik, created a single entity as a single object, denormalised other attributes and temporarily stored the data in memory. xbrldimension.py catch and clean the denormalised data from null, duplicate tuples (a consequence of the denormalisation process), calculate the financial ratios and trends, collect the nonfinancial values, horizontally merge and store them in the database (see figure 14). the financial indicator used in xafr was the most used in a prior study representing the taxpayer profitability, liquidity, solvability and efficiency performance. xbrldimension.py module calculates the logarithmic trend of financial values, ratios and trends from the specified and previous year. the logarithmic trend is selected because of its “honesty” in representing the value fluctuation (see table 2). 31 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 table 2. arithmetic and logarithmic trend comparison. the logarithmic trend is more “honest” in trend fluctuation magnitude and summarisation year value arithmetic trend logarithmic trend 2018 2019 -90.00% -230.26% 2020 500 -50% -69.31% 2021 1000 100% 69.31% xbrl analyzer feature xbrlanalyzer.py module collaborates with ai features-extraction feature in ai framework. the xbrlanalyzer.py module build aggregates pipeline and the view from the stored denormalised entities (see figure 14). xbrl analyzer feature enables financial indicator value and ratio benchmarking by comparing and calculating deviation with its competitor average statistic in a similar industry. fig. 14. xbrl analyzer and ai features-extraction feature xbrl visualizer feature the xbrlvisualizer.py module visualizes the data insight into a chart, diagram, plot and table presented for thorough analysis using an interactive graph from bokeh (ven, 2021) and a rich feature table from data tables (sprymedia, 2021). figure 15 illustrates how the interactive rca analysis can help the tax officer reveal the high-risk taxpayer. b d s wibowo. xbrl open information model for risk based tax audit using… | 32 fig. 15. xbrl visualizer feature sample: interactive risk status in current ratio vs total assets (rca) indicator analysis scatter plot 3.4.2. artificial intelligence (ai) framework ai framework is the core of xafr machine learning classification. ai framework consists of two features: ai features-extractor feature the aifeatures.py module executes the rule-based risk scoring. financial indicators values, trends, and deviations from the average in a similar industry or known standard are used to calculate the risk score (see figure 16). ai machine-learning feature the extracted features by aifeatures.py were treated to eliminate the insignificant features that could reduce the machine learning performance handled by the aimachine.py module. 3.4.3. web framework finally, the calculation and classification result from xbrl and ai framework is presented in the web application handled by web framework. xafr implements model view controller (mvc) pattern for web framework presentation by utilizing flask for web service and mongodb for the data model provider. the webindex.py define all the routes and logic controller function, while the view function would be presented on html files in the templates folder. xafr implement a responsive web view to adapt to any device display dynamically. 4. testing 4.1. unit testing unit testing is essential to ensure each method, class, module, and package works as expected. this report conducts unit tests in two approaches. first, use internal iterative testing in xbrl, ai, and web packages in each module’s primary scope function (see figure 16). 33 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 fig. 16. unit testing in aifeatures.ipynb module’s leading scope example second, use the pytest 6.25 (krekel, 2021) library and create three-unit test modules: test_ai.py, test_xbrl.py and test_web.py (see figure 17) to develop critical checkpoints test for each necessary module functionality. fig. 17. unit testing using pytest 4.2. user acceptance test the user acceptance test (uat) is designed with the story cards and moscow analysis. there are thirty uat for seven-story cards designed. this report conducts uat in two approaches. first, use a tickmark element checking to ensure the thirty uat elements exist and working correctly. the detail is in appendix a. second, using the user survey approach. the survey conducting by demonstrating xafr to the participant for 10 to 15 minutes then each participant was asked to answer the questionnaire for around 15 minutes. the survey used google form (google, 2021a) from 15 participants. the participant is come from various backgrounds and experiences but is still related to tax, finance, and it with multiple ranges of age, gender and experience in using financial analytical tools. the participant demography detail is presented in figure 18. b d s wibowo. xbrl open information model for risk based tax audit using… | 34 fig. 18. participants demography statistic based on the user perspective, in likert scale from 1 – highly unuseful to 7 – highly useful, xafr usability is 6.24 or helpful in overall features and from 1 – highly inaccurate to 7 – highly accurate, xafr accuracy is 5.87 or slightly accurate to accurate in general parts, see table 3 for more detail per feature score. while for the awareness survey, with the scale of 1 for aware, 0 for not sure and -1 for not familiar, the average value is 0.65. from the user perspective, they are 65,6% aware of overall xafr features. table 3. usability and accuracy survey statistic result criteria 1 2 3 4 5 6 7 usability likert scale from highly unuseful (1) to highly useful (7) ten companies diagram risk 6.7% 13.3% 40% 40% ten industries diagram risk 6.7% 6.7% 60% 26.7% risk distribution diagram 7.1% 7.1% 7.1% 50% 28.6% rca (risk, current ratio and asset) diagram 7.1% 14.3% 42.9% 35.7% taxpayer list 6.7% 13.3% 40% 40% taxpayer detail 6.7% 13.3% 46.7% 33.3% xbrl report edgar archive 6.7% 13.3% 13.3% 46.7% 20% benchmark financial indicator diagram 7.1% 7.1% 14.3% 35.7% 35.7% benchmark financial indicator table 6.7% 20% 33.3% 40% risk explanation 6.7% 6.7% 6.7% 46.7% 33.3% financial indicator sorting 6.7% 20% 40% 33.3% machine learning prediction 13.3% 33.3% 20% 33.3% average score 7.1% 6.8% 7.9% 14.6% 41.8% 33.3% 35 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 weighted total score for overall features 6.24 accuracy – likert scale from highly inaccurate (1) to highly accurate (7) risk classification of the blue chips taxpayers 6.7% 26.7% 20% 33.3% 13.3% risk explanation 6.7% 13.3% 13.3% 33.3% 33.3% industries list 6.7% 13.3% 46.7% 33.3% industries risk 6.7% 33.3% 26.7% 33.3% financial indicator sorting 6.7% 13.3% 40% 40% machine learning prediction 26.7% 20% 33.3% 20% average score 6.7% 14.4% 18.8% 35.5% 28.8% weighted total score for overall features 5.87 4.3. system integration test the system integration test is conducted to ensure data integrity is preserved, extracted wholly and correctly, extracted in the correct taxonomy context resulting in a high non-zero value, thus generating a high-quality risk score, explanation, benchmark, and rule-based risk classification. xbrl etl succession rate test this test measures the completeness of data extracted, transformed and loaded to the database. the integrationtest.py testxbrletlsuccessionrate() method handle this test by comparing the total tuples from each collection tsv with the entire valid document stored in the database. the succession rate is 100% for num, pre and sub, but the rate of tag collection is only 91.01% (see figure 19). fig. 19. xbrl etl succession rate test the tag collection can duplicate because it contains the available xbrl taxonomy tag, and when two or more entities have the same event context, they will use an equal tag. hence, the 91.01% etl succession rate is acceptable (see figure 20). b d s wibowo. xbrl open information model for risk based tax audit using… | 36 fig. 20. xbrl dataset to database load process log non-zero value (nzv) extraction rate test this test measures how many non-zero extracted financial indicators are for each taxpayer. if the number of extracted features is 31, the non-zero value rate varies between 0 to 31. the origin of the zero value can come from the objective 0 value or the null value converted to 0. the null value was generated because the algorithm failed to read the taxonomy context. preoptimization, xafr had a poor nzv extraction rate, see figure 21. fig. 21. non-zero-value distribution comparison, before custom dictionary keywords optimisation (left-red border) and after optimisation (right-green border) there are three essential fields to read xbrl taxonomy context: stmt, crdr and tag fields. for example, locating operating income is started by querying the stmt value with “is” for income statement. the crdr value with “c” for credit (normal balance for operating 37 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 income) and tag with regex keyword value indicate operating income. an in-depth evaluation was conducted to improve the rate by comparing the taxpayer that used the reserved accounting standard taxonomy with the custom one and creating a dictionary to capture the keyword for custom taxonomy context optimisation. finally, the nzv extraction rate testing significantly increased using the custom taxonomy context optimisation (see figure 20 and table 4) table 4. nzv extraction rate comparison before and after optimisation condition nzv above threshold entities counts total entities nzv extraction rate before optimization 39 5178 0.6% after optimization 3217 5178 62.1% feature reduction rate test xafr calculate the matrix correlation and display the information in a heatmap diagram. the threshold is ninety per cent. if features correlate equal to or greater than ninety per cent, only one part is preserved (table 5). table 5. high correlation features list, feature with red colour is eliminated feature feature pair correlati on financial_trend_opm financial_trend_gm_opm 94% financial_trend_roa financial_trend_assetturnover 99% financial_benchmark_quickratio financial_deviation_quickratio 92% financial_benchmark_der financial_deviation_der 94% financial_trend_benchmark_ebitm argin financial_trend_ebitmargin 96% financial_trend_benchmark_roa financial_trend_benchmark_assett urnover 97% financial_trend_benchmark_opm financial_trend_gm_opm 90% financial_trend_benchmark_gm financial_trend_gm 94% financial_trend_benchmark_opm financial_trend_opm 95% financial_trend_benchmark_ebitm argin financial_trend_ebitmargin 96% financial_trend_benchmark_npm financial_trend_npm 96% financial_trend_benchmark_basic eps financial_trend_basiceps 95% financial_trend_benchmark_quick ratio financial_trend_quickratio 93% financial_trend_benchmark_curre ntratio financial_trend_currentratio 94% b d s wibowo. xbrl open information model for risk based tax audit using… | 38 financial_trend_benchmark_roa financial_trend_roa 92% financial_trend_benchmark_roe financial_trend_roe 95% financial_trend_benchmark_der financial_trend_der 95% financial_trend_benchmark_roa financial_trend_assetturnover 91% financial_trend_benchmark_asset turnover financial_trend_roa 93% financial_trend_benchmark_asset turnover financial_trend_assetturnover 94% table 6 informs that mostly the industry-level benchmark’s features remain. besides the prior study that reveals the importance of the industry level benchmark for machine learning performance, the other rationale is the random forest feature importance test. while performing 100 iterations to find the best random forest model, the features importance test was also executed. from 100 iterations, it was found that the benchmark features are consistently in the top ten ranks of the most crucial part (see figure 22). fig. 22. random forest feature importance test screenshot target class distribution test xafr classifies the taxpayers into three classes: low risk labelled as 1, medium risk labelled as two and low risk marked as 1. the class name and label is defined based on the risk score calculation and threshold. the risk score threshold should classify the class to mimic the actual distribution in reality as close as possible. after conducting several tests, it was found that the best threshold is the 15% medium risk threshold from -0.15 to 0.15 (see table 6). 39 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 table 6. rule-based risk scoring technique risk score range risk status risk label explanation flag -1 <= x < -0.15 high risk 3 red flag -0.15 <= x <= 0.15 medium risk 2 yellow flag 0.15 < x <= 1 low risk 1 green flag the 15% medium risk threshold will classify every indicator as a medium risk if the standard or average industry level benchmark deviation is between -15% to 15%. for example, if the average net profit margin in a similar industry is 20%, the medium risk is located between 5% to 35%. if less than 5%, then classified as high risk and if greater than 35%, classified as low risk. this formula is the xafr rule-based risk scoring approach (figure 23). fig. 23. target class distribution accuracy score test finally, the accuracy score test is the ultimate test to measure machine learning performance. xafr uses zscore normalisation and hyperparameter to select the best svm and rf model parameters to improve the machine learning accuracy score. the result of the hyperrparameter test is in figure 24 and applied in each classifier model parameter. fig. 24. hyperparameter test result. random forest model (left) and support vector machine (right) b d s wibowo. xbrl open information model for risk based tax audit using… | 40 furthermore, the test to find the best classifier model is conducted by iterating a hundred times for both svm and rf models (figure 25). fig. 25. find best classifier model test the model is dumped into the pickle file at the end of each iteration. while the accuracy score is put to the end of the filename using this structure: classifiertype_modelpreedictor_accuracyscor e.pkl at the end of the test, the best model can be carried out by selecting the pickle file with the highest accuracy score number and using that model for machine learning risk classification prediction (see figure 26). fig. 26. selecting the best classifier model 5. evaluation xafr passes all unit test using two approaches, completed all the user acceptance test and from the survey participant perspective, xafr features is proper and slightly accurate to accurate. xafr features had a high awareness from the survey participant. system integration tests show that xafr has a high xbrl etl succession rate test, and significantly improves the non-zero value (nzv) extraction rate test and found the root cause of the problem. the high feature reduction rate test shows that xafr can recognize and eliminate insignificant features and reduce multicollinearity. the target class distribution test can reveal the best threshold to produce a risk distribution class that mimics the reality as possible, and the accuracy test shows random forest as the best classifier model. xafr can significantly improve the overall performance by enriching the taxonomy context dictionary to recognize more financial indicators, especially for taxpayer that uses custom taxonomy. 6. conclusion and future works conclusion xafr successfully expose the high financial risk taxpayers capable of paying their tax using the rca analysis and sic 41 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 utilization using rule-based risk scoring and machine learning risk classification and can explain the risk classification for risk-based tax audit selection. it can model the latest xbrl oim 1.0 specification processing to generate meaningful information for tax authorities. xafr can process forty million tuples of data to produce insightful information in a minimalist but informative, responsive and interactive display. future works xbrl oim 1.0 dataset contain rich structured and semi-structured meaningful data. it is a treasure for semantic analysis using the natural language processing (nlp) model, providing better risk explanation and risk classification performance. for example, the financial indicator might calculate as medium risk. however, the nlp report analysis found a significant number of sentiment negative words: “fraud”, “loss” which located near words: “exposed”, “mitigated”, “recognized”, “followed”, and “year “then it could be classified as high-risk semantically. furthermore, predicting the range of possible tax debt paid from the specified xbrl financial statements using the regression predictor model is significant for the future development of machine learning in this area. for example, at every start of the year, based on the last taxpayer’s financial indicator, the tax officer could predict using machine learning how much is the tax debt for the recent year if everything is in similar condition. then, at the end of the year, the tax debt prediction is compared with the prediction using the actual financial statement indicator and the actual tax paid. if the difference is significant, it will trigger a warning alarm sent to taxpayer and tax authority. if there is no additional tax payment or clarification from the taxpayer in any regulated period, then the taxpayer is automatically selected for tax audit. references abbasi, a., albrecht, c., vance, a., & hansen, j. (2012). metafraud: a meta-learning framework for detecting financial fraud. mis quarterly, 1293-1327. amazon (2021) cloud services amazon web services (aws), amazon web services, inc. amunategui, m., & roopaei, m. (2018). displaying predictions with google maps on azure. in monetizing machine learning (pp. 195-235). apress, berkeley, ca. anaconda (2021) anaconda with python 3 on 64-bit windows — anaconda documentation. anggia, p. (2019). achieving of income tax with awareness of taxation in indonesia's tax law system. yustisia jurnal hukum, 8(2), 292-308. apache software foundation (2021a) apache hadoop. available at: https://hadoop.apache.org/ (accessed: 31 october 2021). apache software foundation (2021b) apache sparktm unified engine for largescale data analytics. available at: https://spark.apache.org/ (accessed: 31 october 2021). ashtiani, m. n., & raahemi, b. (2021). intelligent fraud detection in financial statements using machine learning and data mining: a systematic literature review. ieee access. b d s wibowo. xbrl open information model for risk based tax audit using… | 42 bootstrap et al. (2021) bootstrap. available at: https://getbootstrap.com/ (accessed: 31 october 2021). canonical (2021) enterprise open source and linux, ubuntu. available at: https://ubuntu.com/ (accessed: 31 october 2021). carroll, j., & morris, d. (2015). agile project management in easy steps. in easy steps. central bureau of statistics of the republic of indonesia (2021) economic and finance publication. central bureau of statistics of the republic of indonesia. available at: https://www.bps.go.id/site/resulttab (accessed: 30 august 2021). el-bannany, m., dehghan, a. h., & khedr, a. m. (2021, march). prediction of financial statement fraud using machine learning techniques in uae. in 2021 18th international multi-conference on systems, signals & devices (ssd) (pp. 649-654). ieee. gmbh (2021) the qr code generator, the qr code generator. available at: https://www.the-qrcode-generator.com/ (accessed: 17 december 2021). gomaa, m. i., markelevich, a., & shaw, l. (2011). introducing xbrl through a financial statement analysis project. journal of accounting education, 29(2-3), 153-173. google (2021a) google forms. available at: https://docs.google.com/forms/ (accessed: 17 december 2021). google (2021b) youtube. available at: https://www.youtube.com/ (accessed: 17 december 2021). hashicorp (2021) vagrant by hashicorp, vagrant by hashicorp. available at: https://www.vagrantup.com/ (accessed: 31 october 2021). hidayattullah, s., surjandari, i., & laoh, e. (2020, october). financial statement fraud detection in indonesia listed companies using machine learning based on meta-heuristic optimization. in 2020 international workshop on big data and information security (iwbis) (pp. 79-84). ieee. hooda, n., bawa, s., & rana, p. s. (2020). optimizing fraudulent firm prediction using ensemble machine learning: a case study of an external audit. applied artificial intelligence, 34(1), 20-30. joblib (2021) joblib: running python functions as pipeline jobs — joblib 1.2.0.dev0 documentation. available at: https://joblib.readthedocs.io/en/latest/ (accessed: 31 october 2021). jurney, r. (2017). agile data science 2.0: building full-stack data analytics applications with spark. " o'reilly media, inc.". khwaja, m. s., awasthi, r., & loeprick, j. (eds.). (2011). risk-based tax audits: approaches and country experiences. world bank publications. kotsiantis, s., & kanellopoulos, d. (2008, november). multi-instance learning for predicting fraudulent financial statements. in 2008 third international conference on convergence and hybrid information technology (1), 448-452. ieee. 43 | international journal of informatics information system and computer engineering 3(1) (2022) 19-44 krekel, h. (2021) pytest: helps you write better programs — pytest documentation. available at: https://docs.pytest.org/en/6.2.x/ (accessed: 31 october 2021). microsoft (2021a) explore windows 11 os, computers, apps, & more microsoft, windows. available at: https://www.microsoft.com/en-gb/windows (accessed: 31 october 2021). microsoft (2021b) visual studio code code editing. redefined. available at: https://code.visualstudio.com/ (accessed: 31 october 2021). mongodb (2021) mongodb: the application data platform, mongodb. available at: https://www.mongodb.com (accessed: 31 october 2021). oracle (2021a) jdk 11. available at: https://openjdk.java.net/projects/jdk/11/ (accessed: 31 october 2021). oracle (2021b) oracle vm virtualbox. available at: https://www.virtualbox.org/ (accessed: 31 october 2021). pallets (2021) welcome to flask — flask documentation (2.0.x). available at: https://flask.palletsprojects.com/en/2.0.x/ (accessed: 31 october 2021). relan, k. (2019). beginning with flask. in building rest apis with flask (pp. 1-26). apress, berkeley, ca. sec, s. (2021) sec, financial statement data sets. available at: https://www.sec.gov/dera/data/financial-statement-data-sets.html (accessed: 25 november 2021). singh, p. (2018). machine learning with pyspark: with natural language processing and recommender systems. apress. sprymedia (2021) datatables table plug-in for jquery. available at: https://datatables.net/ (accessed: 31 october 2021). štěpánek, l., habarta, f., malá, i., & marek, l. (2021, july). “great in, great out” is the new “garbage in, garbage out”: subsampling from data with no response variable using various approaches, including unsupervised learning. in 2021 international conference on computing, computational modelling and applications (iccma) (pp. 122-129). ieee. ven, b.v. de (2021) bokeh. available at: https://bokeh.org/ (accessed: 31 october 2021). venters, c., & mikkilineni, r. (2020, september). representation and evolution of knowledge structures to detect anomalies in financial statements. in 2020 ieee 29th international conference on enabling technologies: infrastructure for collaborative enterprises (wetice) (pp. 58-63). ieee. xbrl international (2021a) an introduction to xbrl, an introduction to xbrl. available at: https://www.xbrl.org/the-standard/what/an-introduction-toxbrl/ (accessed: 7 december 2021). xbrl international (2021b) xbrl & big data, xbrl & big data. available at: https://specifications.xbrl.org/big-data.html (accessed: 28 november 2021). b d s wibowo. xbrl open information model for risk based tax audit using… | 44 xbrl international (2021c) xbrl certified software, xbrl certified software. available at: https://software.xbrl.org/ (accessed: 28 november 2021). yao, j., zhang, j., & wang, l. (2018, may). a financial statement fraud detection model based on hybrid data mining methods. in 2018 international conference on artificial intelligence and big data (icaibd) (pp. 57-61). ieee. 1 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 potential security issues in implementing iaas and paas cloud service models wishnu kusumo agung erlangga*, muhammad rheza ramadhan** *university of glasgow, united kingdom **australia national university, australia *corresponding email: wishnukaerlangga@gmail.com a b s t r a c t s a r t i c l e i n f o as the digital world evolves, so does potential problem that computer users encounter. cybersecurity threats are still evolving and expanding. unfortunately, most computer users do not understand this properly. the cloud models offered by various public cloud providers remain concentrated on infrastructure resources, application platforms, and software services despite the recent increase in the popularity of cloud computing. the first step in this study will be a literature review to get an understanding of accessible cloud service models. the papers chosen for the study spans 2010 to 2020. all data was gathered from pertinent and related literature on cyber security and cloud computing. the following tenets serve as the foundation for this architecture. first, in the described architecture, the perimeter scanner serves as the first entry point for external cyberattacks. firewall and other security layers become next barriers if the attack can get past first layer. on the other side, the machine learning system will detect every successful assault that gets past the security layers. as a result, there are numerous viewpoints and categorization systems for diverse attacks. it is possible to advance cyber security research in the context of cloud technology by merging the results of existing studies and developing international guiding standards. article history: received 18 dec 2022 revised 20 dec 2022 accepted 25 dec 2022 available online 26 dec 2022 aug 2018 __________________ keywords: cloud computing, cloud service model, cyber security, cyber-crime, cloud security framework, machine learning, mitm, mitb, microsoft azure international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 3(2) (2022) 1-20 wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 2 1. background despite the recent explosion in popularity of cloud computing, the cloud models provided by different public cloud providers continue to be focused on three service tiers: infrastructure resources, application platforms, and software services. eventhough, according to coppolino et al. (2017), jang-jaccard & nepal (2014), cyber security dangers are developing and changing. unfortunately, computer users typically do not follow this up with proper comprehension. modest businesses are still often believed to be immune to cracker assaults because of their small size and lack of important data. in the united states, up to 56% of small business owners said they were unconcerned about falling prey to hacking (cnbc, 2021). according to a survey conducted by proofpoint (2019) over millions of tracked cloud user accounts, 72% of cloud users have been the target of villains, which at least one incident had been experienced by 40% of them. in fact, during the same time period, 43% of firms experienced ten or more breaches. according to a poll of 400 it decisionmakers from small and medium-sized enterprises, 74% of workers find it difficult to follow adequate security practises when working remotely (help net security, 2021). unfortunately, researchers have started to see cloud security systems as a cohesive entity comprising the perspectives of cloud vendors and cloud clients, which has attracted them inadequate attention. the private iaas and paas cloud environment are the subject of this project's research. in this study, we tried to answer two questions. first, what effects do the adoption of iaas or paas cloud service models have on cyber security? second, how do researchers learn about iaas and paas security vulnerabilities and threats? to better understand the relationship between cloud service models and cyber security, this study examined those two cloud service models and looked into the consequences of those models' implementation. 2. study context and literature survey virtual machines (vms) are used in cloud computing (microsoft, 2021a) as the virtual equivalent of desktop or laptop. the ability to create many vms connected by a separate virtual network is made possible by cloud computing (vn). according to kaelin (2019; microsoft, 2021b), vn is an isolated and private environment to connect and execute vms, including database applications. it is used to create communication channels between cloud resources, including the on-premises machines of tenants. this study examines two cyberattacks: man-in-the-middle (mitm) and manin-the-browser (mitb). a mitm cyberattack involves an outsider who unwittingly lies to two parties in order to learn or even alter the information shared between them (mallik et al., 2019). based on (owasp, n.d.), mitb employs a similar strategy but adds a trojan horse to learn about and alter the connection between two parties using the target's browser. the following system is used to conduct various experiments linked to this 3 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 research in order to demonstrate some of the critiques that can be made of this paper: 1) mitm assault: arp poisoning as the method, kali linux as the operating system, and ettercap version 0.8.3.1 was the appplication; 2) mitb assault: xss hooking and javascript injection method, kali linux was the operating system as well and the application used was beef version 0.5.3.0. 2.1. cloud service models cloud servers have virtual data centres as one of their features. the same idea governs how cloud computing services work: they offer users on-demand online computing based on their needs. iaas is a cloud model that gives users the ability to create their own computer systems using the resources that have been made available. iaas solutions give businesses the ability to create and maintain computing infrastructure, such as servers, networks, and data storage. this paradigm handles computing resources like memory, data centres, and network storage as a provisioning service (subashini & kavitha, 2011). scalability and infrastructure provision are provided by iaas. paas is a cloud service model that enables users to execute development and deployment utilising any resources made available by the cloud provider. it can be said that iaas is the most basic form of cloud while paas adds runtime environment and an oprtating system to it. therefore, everything that is controlled by the iaas directly underneath paas will always be included. the availability of middleware is another essential requirement for paas. middleware acts as a bridge between the operating system and the programmes it runs. similar to a water pipe connecting two areas to guarantee that water flows smoothly to its destination, it links one platform to another platform. the conceptual alignment between the iaas and paas cloud models is shown in fig. 1. layers of iaas and paas. fig. 1. layers of iaas and paas during the covid-19 pandemic, as online services become increasingly dependable as a means of communication, many cybercrimes target consumers. the emergence of secaas is related to the rising popularity of cloud computing. it provides security in a separate way for cloud operations, data, and applications. additionally, it encourages the use of new security architectures with adaptable and affordable protection measures (hawedi et al., 2018). 2.2. cloud security three components make up a wellknown cyber security model are confidentiality, integrity, and availability (cia). nist's definitions of confidentiality state that maintaining authorised constraints on information access and disclosure, as well as wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 4 safeguards for individual privacy and privileged information, constitutes sustaining confidentiality. integrity encompasses safeguarding against unauthorised data change or loss as well as ensuring the veracity and nonrepudiation of data. the last attribute is availability, which denotes the capability of authorised individuals to access (and modify) data whenever necessary. evangelopoulou provided one illustration of a tiered security system (2021). because it includes a failsafe, layered cyber security makes this method looks like an onion. as seen in fig. 2. cybersecurity defense mechanism with layers, the mechanism being discussed may encompass several forms of defence as can be seen below. fig. 2. cybersecurity defense mechanism with layers however, de donno et al. (2019) put the operating system to iaas while paas and saas were integrated in one layer. they see the network layer as a part of the physical layer rather than a separate entity. according to (campfield, 2021), major cloud service providers like microsoft azure, google cloud platform, and amazon web service (aws) are responsible for protecting the cloud infrastructure, while cloud users are responsible for protecting their own data. while tenants are in charge of maintaining security "in the cloud," they are responsible for protecting users "at the cloud." 2.3. risks to cloud security a number of risks serve as security gaps (almorsy et al., 2010). security issues for cloud models, according to tchifilionova (2010) are not limited to viruses or trojans types. the two researchers' assertion appears to be supported by the numerous types of attacks that take place. table 1 lists cyber security dangers and assaults that have been incorporated into iaas and paas service models from 2010 to 2020 in accordance with the study's purview. the codes started with ja stands for journal articles. table 1. attacks/threats in cloud from journal articles codes journal title and authors types of attacks/threats ja1 an analysis of the cloud computing security problem (almorsy et al., 2010) dos mitm injection attack 5 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 dictionary attack xml-related attack replay attack malware/viruses ddos hypervisor security breach virtual network attack ja2 cloud computing security considerations (a. tripathi & a. mishra, 2011) virtual machine attack hypervisor security breach phishing, fraud, & software vulnerabilities ja3 a survey on security issues in service delivery models of cloud computing (subashini & kavitha, 2011) dos ja4 security issues and solutions in cloud computing (you et al., 2012) data breach data lock-in data remanence virtual machine attack dos ja5 cloud-based ddos attacks and defenses (m. darwish et al., 2013) dos ddos ja6 cloud computing: security and reliability issues (ahamed et al., 2013) dos ddos side channel attack malware/viruses virtualisation leakage insider attack ja7 a survey of security issues for cloud computing (khan, 2016) port scanning data scavenging spoofing virtual machine attack ja8 web services attacks and securitya systematic literature review (mouli & jevitha, 2016) dos spoofing xml injection ja9 a survey on cloud computing security: issues, threats, and solutions (singh et al., 2016) dos data loss data leakage data manipulation ja10 analisis teknik-teknik keamanan pada cloudcomputing dan nebula (future cloud): survey paper (nugraha, 2016) snooping attack dos mitm ja11 data loss data leakage wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 6 data security is the major issue in cloud computing a review (monika & y., 2016) virtual machine attack mitm dos ddos malware ja12 on cloud security attacks: a taxonomy and intrusion detection and prevention as a service (iqbal et al., 2016) side channel attack virtual machine attack phishing mitm cloud malware injection attack ja13 identity and access management in cloud environment: mechanisms and challenges (indu et al., 2018) brute force attack virus/malware virtual machine attack mitm replay attack session/cookie hijacking dos ddos ja14 identifying the top threats in cloud computing and its suggested solutions: a survey (kofahi & al-rabadi, 2018) data loss data breaches account or service hijacking dos ja15 recent security challenges in cloud computing (subramanian & jeyaraj, 2018) snooping attack address spoofing hypervisor security breach side channel attack virtual machine attack ja16 security in cloud computing: a survey (alhenaki et al., 2019) phishing port-scanning attack mitm metadata spoofing attack virtual machine attack ja17 security concerns and countermeasures in cloud computing: a qualitative analysis (anjana & singh, 2019) virtual machine attack shared resources virtual networks hypervisor ja18 cyber-storms come from clouds: security of cloud computing in the iot era (de donno et al., 2019) multi tenancy virtual machine attack mitm ddos injection flaws mitb xml signature element wrapping metadata spoofing 7 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 application-bug level dos ja19 cyber security threats, challenges and defence mechanisms in cloud computing (ahanger & aljumah, 2020) mitm sniffer attack dos ddos connection flooding attack port scanning the categorization of cyber security attacks and threats has remained uneven despite the wealth of literature on cloud security, by the fact that the variety of cyber security risk keeps expanding (jang-jaccard et al., 2014). those who discuss numerous cyber security types of attacks make recommendations to hold against them in several different methods. several options were thoroughly tested (s. zaman et al., 2021). securing a cloud service aims to guarantee that everything created only performs as intended (nunnikhoven, 2021). in order to integrate with the architectural architecture of the cloud, new security techniques must be developed (carlin & curran, 2011). as a result, more study is required to determine the current state of cloud computing security, and this article attempts to fill that gap. 2.4. cloud security using machine learning this study considered azure as the only object. the reason azure was selected is that it offers the most adaptable machine learning (ml) platform to enable the capabilities in building, training, and implementing ml. as seen in fig. 3, nketah (2016) showed to us that microsoft azure also offers mlaas based on a standard workflow. fig. 3. workflow for azure machine learning 3. method in qualitative research, the level of significance of the implications for cyber security incidents are represented in words and analysed through interpretations and categorization (bhandari, 2020). a new insight can be gained by using qualitative methods to explain phenomena, come up with fresh concepts, or develop and test existing hypotheses (fujs et al., 2019). the literary works included for the study span the years 2010 through 2020. in summary, we attempted to analyse the distinctions between iaas and paas based on the body of literature already in existence from the perspective of a cyber-security incidents. to test our methods, we look for criticisms in relevant papers. by developing a novel model with one cloud service provider as a choice, this research also demonstrates the applicability of machine learning (ml) in cloud security. as a result, we concentrated on security wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 8 methods that have been utilised to address security issues in a cloud computing. 4. results and discussion 4.1. secure cloud services as previously indicated, the capabilities and security issues of the cloud model underneath it will be inherited by the cloud service models. we contend that five security levels may be utilised in attacks on the iaas and paas tiers. the virtual layer, physical layer, data storage layer, application layer, and network layer. in terms of their definitions, the five layers are as follows: physical layer, where all the hardware required to run cloud services is located (nist, 2018). virtualization layer, which divides infrastructure resources into virtual computers known as virtual machines (fitzek et al., 2020). data storage layer consists of storage used to store data on the internet. lastly, application layer itself. when viewed from the perspective of the attacker, figs. 4 and 5 demonstrate how the proliferation of attacks in each tier of security. fig. 4. responsibility assumed by the cloud provider for each security layer fig. 5. cloud user’s responsibility percentage for each security layer cloud service providers focus mostly on threats that target the data layer. most attacks on cloud tenants, however, take place at the network layer. given that communication involves networking, which comprises the network interface layer and network access layer, this scenario is plausible (evangelopoulou, 2021). aside from that, cloud tenant is not potentially liable for any physical attacks. after all, they have control over the physical network and infrastructure. as a result, the security of the physical layer is entirely the responsibility of the cloud provider. then, in relation to the overall number of attacks included in this analysis, nearly every form of attack (84.8%) may be the responsibility of the cloud provider. in the meantime, 66.67% of the different sorts of attacks can be the fault of cloud tenants. one of the attack types for which the cloud provider is not liable is the "replay attack." this attack can be executed by monitoring messages delivered or received by cloud tenants. the seized message will then be sent back to the original recipient by the attacker. the recipient will presume that the reply message came from the trusted sender and will accede to the attacker's demands on faith. other forms of attacks, that become cloud users’ responsibility, can 9 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 be lessened in many ways, including the following: (tables 2 and 3) table 2. cloud users’ responsibility types of attacks table 3. lowest occurrence rate of cyber attack based on journal articles the low occurrence rate might lead to a number of different causes. to begin with, cloud tenants rarely experience attacks of this nature. the impact of the attack is also less severe than dos or ddos, which result in physical harm (shutting down the service). third, not many people are aware of the attack's nature. fourth, because attack types are not standardised, it is challenging to group related attack types together. iaas and paas-based cloud services continue to have a wide range of cyber security implications. 4.2. threats and security attacks against iaas and paas cloud models because this service model only covers the most basic services, an iaas cloud provider provides its tenants with the fewest security protections. a saas cloud provider, however, would offer the most services (sosinsky, 2011). iaas users should be aware of the requirement to offer sufficient security for the information and applications they deploy to the cloud. the cloud vendor bears a greater degree of responsibility in the paas model because tenants get a freedom implementing protection for their data in the applications. the operating system they employ serves the same purpose in helping them create their applications. wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 10 when expressing worries about cloud services' security, researchers differ from one another. several of them have made their cyber security flaws public. table 4 shows an attack pattern that spans different cloud security layers. table 4. cia triad indicators of cyber attacks/threats on cloud models attack target location attack types cloud models codes cyber security indicators iaas paas physical dos • • ja1, ja3, ja4, ja5, ja6, ja8, ja9, ja10, ja11, ja13, ja14, ja18, ja19 a ddos • • ja1, ja5, ja6, ja11, ja13, ja18, ja19 a side-channel attack • • ja6, ja12, ja15 c, i, a data storage malware • • ja1, ja6, ja11, ja12, ja13 c, i, a data scavenging • • ja7, ja17 c data loss • • ja9, ja11, ja14 a data leakage • • ja9, ja11 c data breaches • ja4, ja14 c, i data lock-in • • ja4 a data remanence • ja4 c data manipulation • ja9 c virtual virtual machine attack • ja2, ja4, ja6, ja7, ja11, ja12, ja13, ja15, ja16, ja17, ja18 c, i, a virtual network attack • ja1, ja17 c, a hypervisor security breach • ja1, ja2, ja15, ja17 c, a multi tenancy • • ja17, ja18 c, a network xml-related attack • • ja1, ja8, ja18 a snooping attack • ja10, ja15 c port scanning • ja7, ja18, ja19 c sniffer attack • • ja19 c, i, a connection flooding • ja19 c replay attack • • ja1, ja13 c 11 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 mitm • • ja1, ja10, ja11, ja12, ja13, ja16, ja18, ja19 c, i, a application phishing • ja2, ja12, ja16 c spoofing • • ja7, ja8, ja15, ja16, ja18 c, i brute force attack • • ja13 c, i injection attack • • ja1, ja18 c, i, a dictionary attack • • ja1 c, i session hijacking • • ja13, ja14 c, i, a mitb • • ja18d c a. b. nassif et al., (2021) and makkawi & yousif (2020) claim that ddos and dos are the two most common types of cyberattacks. the frequency of incidence of assault kinds is shown in table 4 reveals that the most frequent attacks are virtual machine attack (48%) and dos (71%). this section tries to evaluate the data in table 4. and to observe and describe how cyber security threats in the iaas and paas model clouds have increasing and decreasing effects. the cia triad in the given figure shows their impact on the cyber security layer. fig. 6 shows the effects of cyber-attacks as mentioned in table 4. fig. 6. effect of attacks on security layers. as depicted in fig. 6, the application level is where the most substantial effects of cyberattacks in the iaas and paas cloud service models occur. such attacks can be mitigated or prevented by several proactive measures, including robust authentication standards, hashing technology, encryption, or application updates. while only 3.8% of all current cyberattacks have a physical component, this is where confidentiality is least affected. the likelihood of it occurring is high since cyberattacks on the physical infrastructure of public cloud services will only take place on the premises of cloud service providers. additionally, the physical restrictions on the cloud provider's authorization of access privileges to the cloud centre are very severe. for instance, google has six layers of tight security to prevent unauthorised access to the google cloud data center area, starting with the most basic precautions like signage and fencing and moving up to smart fencing and thermal cameras on the second tier (google cloud tech, 2020). wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 12 attacks and threats on the cia triad component in fig. 6 have varying degrees of impact. for instance, integrity at the data and network levels is where the lowest implication values are derived. the highest value for both levels has effects on the same side, notably confidentiality, at the same time. due to a person's intense desire to discover the secret of confidential material, one might hypothesise that a high score on confidentiality is most often the result of hacking or social engineering, posing the risk of insufficient authentication or authorization. the high number of attacks threatening the availability of virtualization are undoubtedly influenced by growing network and device deployment, as well as developer ignorance of the hypervisor code. a single management console's efficient management, monitoring, and configuration of vms and vns can lessen the risks they pose (a. tripathi & a. mishra, 2011). from table 1, we discovered fascinating information on the researchers' classification of the various cyberattacks. there is a lack of standardisation in the name and categorization of cyber-attacks when looking at the sorts of attacks in the researchers' findings above. for instance, some researchers include mitm kind of attack as a component of service-oriented architecture (almorsy et al., 2010; iqbal et al., 2016). while (indu et al., 2018) put this attack into human-centered security due to an insecure ssl architecture. interestingly, (de donno et al., 2019) took this issue as a result of a lack of network security. even though the methodology used in mitm attacks and mitb attacks is the same (owasp, n.d.), de donno et al. (2019) distinguish between those two methods and classify mitm in network security level. interesting distinction between the two is made by the method used. mitb can use trojans like zeus when launching attacks. while mitm will employ a physical technique to access a wi-fi router, such as in public area, houses without sufficient security measures in place, or websites without an ssl certificate configured (use https). to demonstrate that these two attacks operate at the same security level, the analysis below uses attacking experiments with mitm and mitb applications. 1) mitm ip target : 10.0.2.2 target url : http://libgen.li method : mitm and arp poisoning tool : ettercap version 0.8.3.1 an illustration of the mitm attack is shown in figs 7 and 8. the web server and the web browser connection is not protected when users check in to sites like http://libgen.li without an ssl certificate, making it possible for attackers to intercept sensitive data like usernames and passwords. 13 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 fig. 7. display of the site login page at libgen.li fig. 8. successful mitm method login information capture the possible risks that mitm provides to targets are demonstrated in figs. 7 and 8. the victim of the assault provided a username and password to htpp:/libgen.li using the arp poisoning technique, which the mitm that was performed was able to retrieve using the ettercap attack programme version 0.8.3.1. the mitm method may be used to attack cloud tenants with a more advanced strategy. 1) mitb google chrome is the preferred browser method : xss hooking with javascript injection tool : beef version 0.5.3.0 fig. 9. shows an illustration of the mitb attack pattern. unfortunately, in the experiment, mitb assaults could not be launched using the zeus trojan. on the other hand, the hook.js trojan, which is utilised as a pilot in the beef application, might be used to conduct the mitb attack simulation. figs. 9-11 show the simulation of the mitb assault. fig. 9. map of mitb attack using hooked browser and beef fig. 9 shows the position of the attackers in mitb. the hook.js trojan was successfully installed by the attacker (beef) on the target's google chrome browser. the attacker could continue their operations after successfully hijacking the chrome browser. as an example, as seen in fig. 10, it was done by spoofing the target's google chrome browser with a google mail login page. wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 14 fig. 10. webpage phishing on mitb attack the target's information in the aforementioned example is username: anythinguwant and password: tampan100. as seen in fig. 11, the credential used could be noted in the beef application. fig. 11. information retained on the infected chrome browser via fig. 11, mitb noted that the username and password used to attack the target were exactly the same as those used by the victim. as a result, the two attacks (mitm and mitb) may have a similar structure. the inference that may be made is that attacks like mitm and mitb can be executed at the same level, by taking advantage of network security flaws. the vulnerability results in packets being sent over the network, giving attackers access to view data traffic. then, as was already noted, the replay attack by uses the same attack strategy as mitm. it entails the presence of undesirable people in the communication path between the sender and the recipient. however, due to the lack of standardised knowledge of the consequences of cyber security, some people now view it as of human-centered security issue (indu et al., 2018) while almorsy et al. (2010), saw it from the perspective of network security. in addition, dos and xml attacks are distinguished by almorsy et al. (2010). however, a closer look reveals that dos is actually an assault on the xml parser. therefore, we contend that these two ideas can be combined into one section, namely dos. since the beginning of the relationship, the cloud service provider has been responsible for data security. as a result, tenants will be subject to local and international laws that are in force in the countries where the cloud service provider stores their data. however, if a data leak occurs while a tenant is using cloud computing services provided by a cloud provider, the tenant could still be held accountable for the data loss. a service that handles everything from physical security to environmental security to virtualization security is ec2, for instance. tenants are in charge of all applicationand data-related problems. the idea to include security as a form of cloud computing service (forcepoint, 2019; hawedi et al., 2018) seems like it could advance given the significance of cyber security risks. 4.3. cloud security as a service the security of paas and saas will also be impacted in the event of an attack on the iaas platform (hashizume et al., 2013). these interdependencies put these systems' security in jeopardy. solution offered by hawedi et al. (2018) focuses on providing a security service that enables cloud tenants to keep an eye on their virtual machines (vms) by putting in place a security mechanism that is tailored to their specific requirements. 15 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 the results of the initial examination show that researchers still lack a comprehensive manual on cloud service dangers and attacks. using alarm signals as warnings, all forms of cloud system. security services halt till the investigation stage. therefore, the policies chosen by cloud tenants will dictate the subsequent actions. there is a danger involved in regarding secrecy or confidentiality. data loss or leakage may occur while cloud tenants are deliberating how to proceed. in accordance with this, an ml system could be created as a further development in this situation. 4.4. technology & tools for cloud computing security in terms of security, more consideration should be given to cloud development. several cloud services companies, including amazon, google, and microsoft, presently offer mlaas. the algorithm is capable of overcoming the attack pattern. thus, the subsequent attack pattern will guide cloud security systems to function successfully. as a result, when the model is deployed, it can be used appropriately. microsoft azure offers a variety of algorithms that users can carry out the most suitable predictive analytics tasks. cyber threats, however, are usually more difficult to anticipate and have no real economic value. it is possible, though, for a deep learning framework to measure the risk of cyberattacks and to handle them precisely. a summary of the calibre of each ml technology provided is given through comparisons ever done by the three biggest cloud providers (amazon, azure, and google). even so, the assessment considers a few factors, including the mode of operation, the forecast result, the data processing time, and the algorithm, various datasets can yield different results (nketah, 2016). 4.4.1. proposed design azure web application firewall, azure ad identity protection, azure ddos protection are just a few of the choices microsoft azure offers to increase the security of tenant-run cloud systems. even connecting to amazon web services is possible with some of them. predictive analytics can be developed as an extra autonomous cloud security system to identify internal and external assaults using ml technology operating within azure security. the tenants themselves can put together the necessary algorithm to produce a desired prediction model if ml can correctly carry out the training, validation, and testing processes. we suggest taking a number of actions to support the adoption of predictive analytics by renters as a remedy by carrying out a number of procedures as follows: 1) select one-class svm or pca-based anomaly detection as your ml algorithm model for anomaly detection; 2) establish the criteria that will be used to classify the assault type. it can also serve as a framework for cloud security implementation; 3) based on the model of choice, set the predefined parameters; 4) run a preliminary supervised ml model to identify anomalies brought on by threats or attacks; 5) verify the output of the suggested ml model; wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 16 because it's possible that a brand-new event happens that wasn't brought on by an attack or threat, but rather something like a brand-new service. in the meantime, the attack must be retaliated to as soon as feasible. however, the disorganised system of categorising cyberattacks to offer the best defence against attacks has not yet attained standardisation. therefore, we propose that the taxonomy issue should be ignored. we suggest an ml security detection model for azure that works at both the iaas and paas levels. the suggested model distinguishes between internal and external threats. for an internal threat, the model combines machine learning (ml) models at the first layer and classifies it based on a number of criteria. furthermore, an external attack will target the second layer ml because it is more likely to be seen than an internal threat. fig. 12 illustrates the model’s conceptual design. fig. 12. proposed combined ml framework for microsoft azure cloud service security the following tenets serve as the foundation for this architecture. first, in the above-described architecture, the perimeter scanner serves as the first entry point for external cyberattacks. the firewall serves as a barrier if the attack manages to get past the perimeter scanner and so does the other security layer. on the other side, every successful assault that gets past the security layers will be detected by the machine learning system. a speedy and accurate response to anomalies brought on by successful external attacks is made possible by the by the anomaly detection machine learning algorithm. the multiclass classification algorithm will be used to reenter each of these anomalies. this system will identify and look into any unusual behaviour by applying the categorization principle. as a result, if the azure defender system's machine learning design for security is sufficiently matured, third-party services to offer additional cloud security services may not be required. 5. conclusion cyber-attack types and techniques are changing. unfortunately, there is no shared vocabulary for interpreting this in the cyber security frameworks until today. as a result, there are numerous viewpoints and categorization systems for diverse attacks. it is possible to enhance cyber security studies in the context of cloud technology by fusing the results of that study and developing a single international guiding standard. the responses to the research questions discovered by this study can be summed up as follows: 1. iaas and paas cloud service model implications for cyber security are still extremely extensive. integrity impacts from cyberattacks are the most serious consequences and happen at the application level. the lowest implication, which only applies to the iaas cloud service model, occurs at the physical level in confidentiality. 2. 17 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 researchers categorised several attack methods in non-standard cloud service models. however, many people do not differentiate between different cloud service levels (iaas and paas). again, this can be troublesome because iaas and paas run on different tiers. subjectivity is one of the methods used in this paper's. researchers have used a variety of techniques to explain the effects of cyber threats on iaas and paas cloud services as a result of this situation. to standardise and streamline the terminology used to describe attacks, several researchers have created a taxonomy for cloud attacks. nevertheless, the repercussions of cyberattacks are highly diverse. mlaas comes as one solution for cloud model security system. especially for iaas and paas cloud computing levels. research suggestions two of the most significant obstacles to this research is data collection and ml model training to show that mlaas may become truly possible to run. there must be several distinct and useful attacks carried out to see how ml operates and can perform predictive analytics. the proposed model will be used in subsequent studies to perform trustworthy ml testing. moreover, since unsw-nb15 represents contemporary low footprint attacks, guaranteed data sets like these can be chosen as training data in finding the highest rate of successful ml algorithm. references ahamed, f., shahrestani, s., & ginige, a. (2013). cloud computing: security and reliability issues. communications of the ibima, 2013, 1. alhenaki, l., alwatban, a., alahmri, b., & alarifi, n. (2019). security in cloud computing: a survey. international journal of computer science and information security (ijcsis), 17(4). aljumah, a., & ahanger, t. a. (2020). cyber security threats, challenges and defence mechanisms in cloud computing. iet communications, 14(7), 1185-1191. bhandari, p. (2020). a step-by-step guide to data collection. scribbr. https://www.scribbr.com/methodology/data-collection/ campfield, m. (2021). mind the gap: the cloud security skills shortage. computer fraud & security, 2021(8), 6-10. coppolino, l., d’antonio, s., mazzeo, g., & romano, l. (2017). cloud security: emerging threats and current solutions. computers & electrical engineering, 59, 126-140. curran, k., carlin, s., & adams, m. (2011). cloud computing security. journal of network engineering, 37(1), 4069-4072. darwish, m., ouda, a., & capretz, l. f. (2013). cloud-based ddos attacks and defenses. in international conference on information society (i-society 2013) (pp. 67-71). ieee. wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 18 de donno, m., giaretta, a., dragoni, n., bucchiarone, a., & mazzara, m. (2019). cyber-storms come from clouds: security of cloud computing in the iot era. future internet, 11(6), 127. evangelopoulou, m. (2021). cyber security fundamentals (m). fitzek, f. h. p., granelli, f., & seeling, p. (2020). network slicing. in computing in communication networks (1st ed., p. 71). academic press. forcepoint. (2019). what is security as a service (secaas)? forcepoint. https://www.forcepoint.com/cyber-edu/security-as-a-service-secaas fujs, d., mihelič, a., & vrhovec, s. l. (2019, august). the power of interpretation: qualitative methods in cybersecurity research. in proceedings of the 14th international conference on availability, reliability and security (pp. 1-10). google cloud tech. (2020). google data center security: 6 layers deep. https://www.youtube.com/watch?v=kd33uvzhnaa hashizume, k., rosado, d. g., fernández-medina, e., & fernandez, e. b. (2013). an analysis of security issues for cloud computing. j internet serv appl 4 (1): 1– 13. hawedi, m., talhi, c., & boucheneb, h. (2018). security as a service for public cloud tenants (saas). procedia computer science, 130, 1025-1030. help net security. (2021). what is the impact of remote work on security best practices? help net security. https://www.helpnetsecurity.com/2021/06/24/remote-work-securitypractices/ indu, i., anand, p. r., & bhaskar, v. (2018). identity and access management in cloud environment: mechanisms and challenges. engineering science and technology, an international journal, 21(4), 574-588. iqbal, s., kiah, m. l. m., dhaghighi, b., hussain, m., khan, s., khan, m. k., & choo, k. k. r. (2016). on cloud security attacks: a taxonomy and intrusion detection and prevention as a service. journal of network and computer applications, 74, 98-120. jang-jaccard, j., & nepal, s. (2014). a survey of emerging threats in cybersecurity. journal of computer and system sciences, 80(5), 973-993. jensen, m., schwenk, j., gruschka, n., & iacono, l. l. (2009, september). on technical security issues in cloud computing. in 2009 ieee international conference on cloud computing (pp. 109-116). ieee. kaelin, m. (2019). how to create a cloud-based virtual network in microsoft azure. techrepublic. https://www.techrepublic.com/article/how-to-create-acloud-based-virtual-network-in-microsoft-azure/ khan, m. a. (2016). a survey of security issues for cloud computing. journal of network and computer applications, 71, 11-29. kofahi, n. a., & al-rabadi, a. r. (2018). identifying the top threats in cloud computing and its suggested solutions: a survey. networks, 6(1), 1-13. 19 | international journal of informatics information system and computer engineering 3(2) (2022) 1-20 makkawi, a. m., & yousif, a. (2020). machine learning for cloud ddos attack detection: a systematic review. in 2020 international conference on computer, control, electrical, and electronics engineering (iccceee) (pp. 1-9). ieee. mallik, a. (2019). man-in-the-middle-attack: understanding in simple words. cyberspace: jurnal pendidikan teknologi informasi, 2(2), 109-134. microsoft. (2021a). virtual network – virtual private cloud | microsoft azure. virtual network. https://azure.microsoft.com/en-gb/services/virtual-network/ microsoft. (2021b). what is a virtual machine and how does it work | microsoft azure. https://azure.microsoft.com/en-gb/overview/what-is-a-virtual-machine/ monika, g., & kalpana, y. (2016). data security is the major issue in cloud computing-a review. indian journal of science and technology, 9, 43. mouli, v. r., & jevitha, k. p. (2016). web services attacks and security-a systematic literature review. procedia computer science, 93, 870-877. nassif, a. b., talib, m. a., nasir, q., albadani, h., & dakalbab, f. m. (2021). machine learning for cloud security: a systematic review. ieee access, 9, 20717-20735. nist. (2018). evaluation of cloud computing services based on nist sp 800-145. https://www.nist.gov/publications/evaluation-cloud-computingservices-based-nist-sp-800-145 nketah, g. u. (2016). comparison of machine learning services. university of stavanger, norway. http://hdl.handle.net/11250/2413901 nugraha, b. (2016). analisis teknik-teknik keamanan pada future cloud computing vs current cloud computing: survey paper. jurnal nasional teknologi dan sistem informasi, 2(2), 35-42. nunnikhoven, m. (2021). top cloud security challenges for 2021. trend micro. https://www.trendmicro.com/en_se/devops/21/b/top-cloud-securitychallenges-for-2021.html owasp. (n.d.). man-in-the-browser software attack | owasp foundation. retrieved november 15, 2021, from https://owasp.org/wwwcommunity/attacks/man-in-the-browser_attack singh, a. (2019). security concerns and countermeasures in cloud computing: a qualitative analysis. international journal of information technology, 11(4), 683690. singh, s., & jeong, y. s. park, jong.(2016). a survey on cloud computing security: issues, threats, and solutions. journal of network and computer applications, 75. sosinsky, b. (2011). cloud computing bible. wiley publishing, inc. subashini, s., & kavitha, v. (2011). review: a survey on security issues in service delivery models of cloud computing. j. netw. comput. appl, 341. subramanian, n., & jeyaraj, a. (2018). recent security challenges in cloud computing. computers & electrical engineering, 71, 28-42. tchifilionova, v. (2010). security and privacy implications of cloud computing–lost in the cloud. in international workshop on open problems in network security (pp. 149-158). springer, berlin, heidelberg. wishnu kusumo agung erlangga. cloud service model and its cyber security implications | 20 tripathi, a., & mishra, a. (2011). cloud computing security considerations. in 2011 ieee international conference on signal processing, communications and computing (icspcc) (pp. 1-5). ieee. wang, x., zhang, y., zhang, h., wei, x., & wang, g. (2019). identification and authentication for wireless transmission security based on rf-dna fingerprint. eurasip journal on wireless communications and networking, 2019(1), 1-12. you, p., peng, y., liu, w., & xue, s. (2012). security issues and solutions in cloud computing. in 2012 32nd international conference on distributed computing systems workshops (pp. 573-577). ieee. zaman, s., alhazmi, k., aseeri, m. a., ahmed, m. r., khan, r. t., kaiser, m. s., & mahmud, m. (2021). security threats and artificial intelligence based countermeasures for internet of things networks: a comprehensive survey. ieee access, 9, 94668-94690. 1 | international journal of informatics information system and computer engineering 2(1) (2021) 15-24 implementation of information system in indonesian traditional beverage businesses hadi purnomo *, f r fitrah **, r maulana***, m m pratadina**** *departemen manajemen, universitas komputer indonesia, indonesia departemen sistem informasi, universitas komputer indonesia, indonesia e-mail: * hadi.purnomo@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o this study aims to apply the information system to every business owner who can inform businesses beverage container business through information technology. the function of this information system's application is for business owners and customers interested in buying products business. this study is an application of information systems in the indonesian traditional beverage business. this study's method was conducted using qualitative descriptive writing method of collecting data through literature study and observation to the growing beverage business owners. information system development is conducted using the waterfall method. the results of this study show an overview of how the information system is expected by the business owner as well as the beverage business impact that can provide the quality of the business product beverage business is being run. the conclusion is that there is a large information container such as a website application applied to the indonesian traditional beverage business so that website visitors can see, buy, or provide suggestions or comments regarding the already available products article history: ___________________ keywords: information system, business, beverage, traditional, international journal of informatics information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 2(1) (2021) 15-24 received 8 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 purnomo et al. implementation of information system in indonesian ...| 16 1. introduction indonesia is a local cultural heritage with its respective functions included with traditional food and beverage. there are natural ingredients such as ginger, turmeric, and others (soegoto, et al., 2018). the information system is part of the information technology that includes all devices such as computers, software, internet, databases, communications systems, mobile devices, and others. the information system is very thick with the actor who plays as well as with other organizational (alter, s. 2013). therefore, the application of information systems in the traditional indonesian beverage business is needed as a container of information for many people. the majority of visitors come to indonesia for the first time, as much as (78%) know about food and drink indonesia before they come to indonesia. the tourists know indonesian culture because getting information directly from friends, family, or other woods is more aware than hearing it straight from newspapers and articles about indonesia's food and beverage (wijaya, et al., 2016). indonesian food and drink are described as a tree because of indonesia's diversity of ethnic groups and cultures. this point can be illustrated by a matrix of how much indonesian food and drink is found in some areas that are frequently or rarely visited (situngkir, et al., 2015). based on data from the department of industry and trade of east java in 2007, there were 617 more smes processed food and beverage products in indonesia. the data were selected from the three counties or cities that have a majority of umkm, namely surabaya, malang, and kediri (wanto, h. s. 2012). the research in jakarta on indonesia's processed food and beverages is divided into six regions: jakarta central, north, south, east, west, and kepualan seribu. only five cities have been invested by the international brands of fast-food restaurants (widaningrum, et al., 2018). the system design of expert traditional drink recipes implementation case-based reasoning is a web-based application. the user can see complete drink recipe information as well as search for drink recipes by inputting ingredients. the system will search the data in the database beverage drinks menu (darmawan, r., & wibisono, s. 2019). several sectors of the indonesian processed food and beverage industry are experiencing high growth. this industry's existence is enormous opportunities; the future food industry and beverage indonesia become a mainstay for indonesia's economic progress (widodo, s. 2019). preserving agricultural products by fermentation has been carried out since the early seventeenth century, especially for tempe. local and traditional wisdom shows spontaneous fermentation, mostly involving mixed cultures, including lactic acid bacteria, conducting biopreservation of perishable agricultural products (surono, i. s. 2016). the purpose of this paper is for beverage business owners to be able to monitor the process of providing information about their business and customers who will see and buy beverage products. the method used for this paper is the observation and literature study method system development used the waterfall (van casteren, w. 2017). 2. method in this research, the method used is literature study and observation, by 17 | international journal of informatics information system and computer engineering 2(1) (2021) 15-24 visiting the beverage business place, observing and collecting data from business owners. in addition, information system development is conducted using the waterfall method (see fig. 1). 3. results and discussion in this study, we created a system where the system can be implemented into a web application that serves as a content provider of indonesian traditional beverage product introductions. 3.1. procedure the procedure is the initial stage in the manufacture of an information system implemented into a web application. where the procedure contains an explanation of the system flow web will be created. 1. the user enters the traditional drinks web application's home page to see explanations and pictures of traditional indonesian drinks. 2. after that user can see the menus available in the web application that consists of a homepage, beverage products, and admin contacts. 3. then the user can click on the menu of beverage products that contain traditional drinks originating from indonesia. 4. users can view each traditional indonesian beverage product's details by clicking on one of the images available on that page. 5. then the user will go to a page that contains details of the drinks he chose 6. after that, the user can select the contact menu containing the admin's bio and contacts for the web application. 7. on this page, the user can also provide feedback and rating suggestions to the web application admin. 8. after that, the admin can update and add data to the list of traditional indonesian beverages. 3.2. flow map flowmap is a depiction of an ongoing activity that shows the movement of data that serves as an explanation in a flow of procedures. flowmap in the form of images that are linked to each other so that it becomes an information system that functions as an overview of the programmer to create the web (see fig. 2). purnomo et al. implementation of information system in indonesian ...| 18 fig. 1. the waterfall model fig. 2. flowmap 19 | international journal of informatics information system and computer engineering 2(1) (2021) 15-24 3.3. context diagram diagram context is the design of functions consisting of several diagrams that are used to see the correspondence between the program and the code created. diagram context can also be referred to as the earliest diagram or data flow diagram (dfd) level 0 and will be continued at dfd level 1 which will be more detailed about the flow diagram (see fig. 3). 3.4. dfd level 1 a data flow diagram (dfd) is a data flow used in system design. in this case, there are three symbols; entities, process, and file/database. it also has two entities are user and admin. in addition, the systems also have one file/database and one process namely main menu. everything will be interconnected and the systems will input data into the existing file or database (see fig. 4). fig. 3. context diagram fig. 4. dfd level 1 purnomo et al. implementation of information system in indonesian ...| 20 3.5. mock up a mockup is early in the website's design development, where the mockup serves as a visual design made to resemble the original form of the website to be built. therefore, programmers have a design concept in building a website. where in making this website design mockup consists of five website pages. in addition, each page has its function. home page fig. 5 shows a home page. in this page, all of the information will be presented. there are three sub-menus on the home page, namely home, product, and contact. product page fig. 6 shows the product page. on this page, all of the information regarding the products will be presented. this page also shows the list of the products. product detail page fig. 7 shows a product detail page. on this page, the information regarding the product details will be presented. contact fig. 8 shows the contact page. on this page, it is provided a direct contact such as whatsapp or email of the contact person. it is presented for the user to contact the whatsapp number or email listed if they are interested in the products. feedback and rating page fig. 9 shows a feedback and rating page. on this page, the user can give their feedbacks and ratings regarding the products and website. fig. 5. home page 21 | international journal of informatics information system and computer engineering 2(1) (2021) 15-24 fig. 6. products page fig. 7. products detail page purnomo et al. implementation of information system in indonesian ...| 22 fig. 8. contact page fig. 9. feedback and rating page 23 | international journal of informatics information system and computer engineering 2(1) (2021) 15-24 4. conclusion the conclusion from the results obtained is that a platform such as a website application has been applied to the indonesian traditional beverage business. it is applied so that website visitors can see, buy, or provide suggestions or comments regarding the already available products. references alter, s. (2013). work system theory: overview of core concepts, extensions, and challenges for the future. journal of the association for information systems, 72. darmawan, r., & wibisono, s. (2019). siastem pakar resep minuman tradisional menggunakan metode case-based reasoning. situngkir, h., maulana, a., & mauludy, r. (2015). a portrait of diversity in indonesian traditional cuisine. available at ssrn 2703706. soegoto, e. s., azhari, r. h. g., & istiqomah, a. o. (2018). development of desktopbased information system on waste management bank. in iop conference series: materials science and engineering, 407(1), p. 012058. surono, i. s. (2016). ethnic fermented foods and beverages of indonesia. in ethnic fermented foods and alcoholic beverages of asia (pp. 341-382). springer, new delhi. van casteren, w. (2017). the waterfall model and the agile methodologies: a comparison by project characteristics. research gate, pp. 1-6. wanto, h. s. (2012). the effect of organizational culture and organizational learning towards the competitive strategy and company performance (case study of east java smes in indonesia: food and beverage industry). information management and business review, 4(9), pp. 467-476. widaningrum, d. l., surjandari, i., & arymurthy, a. m. (2018). visualization of fastfood restaurant location using geographical information system. in iop conference series earth environment science. widodo, s. (2019). strategy of strengthening food and beverage industry in indonesia. journal of economics and behavioral studies, 11(4 (j)), pp. 102-110. purnomo et al. implementation of information system in indonesian ...| 24 wijaya, s., morrison, a., nguyen, t. h., & king, b. (2016). exploration of culinary tourism in indonesia: what do the international visitors expect? in asia tourism forum 2016-the 12th biennial conference of hospitality and tourism industry in asia. atlantis press. 1 | international journal of informatics information system and computer engineering 1 (2020) 13-22 enhanced the weighted centroid localization algorithm based on received strength signal in indoor wireless sensor network medhav kumar goonjur*, irfan dwiguna sumitraᶲ, sri supatmiᶲ *orange business services (mauritius) ltd, mauritius ᶲmagister sistem informasi, universitas komputer indonesia, indonesia a b s t r a c t s a r t i c l e i n f o a challenging problem that arises in the wireless sensor network (wsn) is localization. it is essential for applications that need information about target positions, are inside an indoor environment. the localization scheme presented in this experiment consists of four anchor nodes that change their position coordinates and one target node that is used to control the distance. the localization algorithm designed in this paper makes use of the combination of two algorithms; the received strength signal indication (rssi) and weight centroid localization algorithm (wcla), called the rssi-wcla algorithm. the laboratory results show that the fusion between the rssi-wcla algorithm is outstanding than rssi and wcla algorithms itself in terms of localization accuracy. however, our proposed algorithm shows that the maximum error distance is less than 0.096m. article history: received 3 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 e online 09 sep 18 ___________________ keywords: indoor localization, rssi, wcla, wsn. international journal of informatics information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 1 (2020) 13-22 goonjur et al. enhanced the weighted centroid localization algorithm based on … –.| 14 1. introduction research on an indoor localization algorithm is a challenging problem due to the various number of complexities that arises in an indoor environment condition, including counting the nearness of impediments such as dividers, entryways, furniture, individuals within the room, and other components. furthermore, indoor localization is more enticing and getting more attention from the research domain where gps signal is not available to detect accurate positions in the building. in order to provide seamless localization and positioning service for indoor environment, some approaches in addition to wireless sensor networks (wsn) have been successfully proposed to arrange accurate solutions as the primary key for location-based service (lbs) (zhang, s., & xing, t., 2013). addressing some of these issues, we can calculate the estimated positions using the distance of both the transmitter and the receiver that is known in the rangebased scheme, among some of the conventional distance measurement algorithms, such as the aoa (angle of arrival), tdoa (time difference of arrival), toa (time of arrival), and rssi (received strength signal indication). rssi method to begin with employments the receiver nodes to recognize lost signal throughout communication, and calculate the positioning using at least three nodes as a beacon node, which already know of their coordinates. rssi method does not necessarily need any additional hardware device or extra power, but communication bandwidth is required to calculate rssi (mukhopadhyay, b., et al, 2014). toa calculates the signal of time through the transmitter to the receiver in the sensor node field. however, this method has drawbacks when using a ranging approach, and between the two nodes should be synchronized in precise clock. the time difference of arrival between the arriving node and the transmitter was measures by tdoa. for accuracy of time difference, the method is implemented to discover broadband signal. measuring the time difference of arrival at a particular array element used aoa. it is capable of a narrowband signal. aoa needs other equipment to estimate arrival angel in wireless signals (suzhe, w., & yong, l., 2012). on the other hand, a range-free scheme also a popular algorithm for localization. according to network connectivity, these schemes can obtain the nodes location information. one of these methods is the classical centroid localization algorithm (cla) that was proposed by (chen, m., & liu, h., 2012). in this algorithm, all localized node estimate the centroid of all obtained beacon’s positions. however, cla inherent bias is still significant. it can be outperformed by weight centroid localization algorithm (wcla), which uses the received strength signal to quantify the anchor nodes in range and emphasizes nearer ones (behnke, r., & timmermann, d., 2008). on the other hand, a range-free scheme also a popular algorithm for localization. according to network connectivity, these schemes can obtain the nodes location information. one of these methods is the classical centroid localization algorithm (cla) that was proposed by (chen, m., & liu, h., 2012). in this algorithm, all localized node estimate the centroid of all obtained beacon’s positions. however, cla inherent bias is still significant. it can be outperformed by weight centroid localization algorithm (wcla), which 15 | international journal of informatics information system and computer engineering 1 (2020) 13-22 uses the received strength signal to quantify the anchor nodes in range and emphasizes nearer ones (behnke, r., & timmermann, d., 2008). according to the pro and cons of two sorts of localization algorithms, this paper proposed a fusion between rssi and wcla in order to increase the estimated accuracy of the distance between target and anchor nodes in an indoor environment. we analyze the impact of a parameter within the rssi to separate estimation. at that point, propose the framework demonstrate utilizing rssi-wcla, which is more convenient, lightweight, and practical. the comparison among these algorithms are shown in table 1. there are six basic indoor positioning algorithms: proximity, aoa, toa/tdoa, rss, dead reckoning, as well as map matching. each of these algorithms have their own pro and cons to address the accuracy problem. in section ii, we presented rssi, cla, and wcla, which construct the fundamental of our proposed algorithm. the experimental setup is explained in section iii, and after that section, iv has appeared the analysis of the test comes about. at long last, the conclusions are given within the final section. 2. localization algorithm model 2.1. rssi based on radio propagation channel some radio propagation on the path loss models is commonly used, namely log-distance path loss model, free space propagation model, log-distance distribution model, hata model, and others. it is utilized because of multipath, fading, shadowing, and timely variation in the environment. the observed model used the shadowing model, which is used in wireless signal transmission. the relationship between separate and gotten control can be communicated concurring to the taking after equation: [𝑃𝑙 (𝑑)] = [(𝑃𝑙 (𝑑0)]𝑑𝐵𝑚 − 10𝑛 log ( 𝑑𝑚 𝑑0 ) + 𝑋𝑑𝐵𝑚 (1) in this model, 𝑃𝑙 (𝑑) signify the loss of path at partition distance set in db relatively to 1 mw. 𝑃𝑟 (𝑑0) acted as the path loss for the reference distance. 𝑑0 is the reference distance which breaks even with 1 meter, dm is the distance between transmitter and receiver in meter, 𝑋𝑑𝐵𝑚 is a zero-mean gaussian distributed random variable and the mean is zero. the standard deviation is commonly for 4-10. these values symbolize the strength signal changed that is obtained in a specific distance; the esteem of n depends on the particular propagation environment. comparatively, n is equal to 2 in a free space environment. in addition, n has tremendous value when obstructions are present. thus, it obtains the shadowing equation model, as shown in equation (2). [pr(𝑑)] = [𝑃𝑟 (𝑑0 )]𝑑𝐵𝑚 − 10𝑛 log10 ( 𝑑𝑚 𝑑0 ) (2) where, 𝑑0=1m, therefore the equation of distance measurement according to rssi value is utilized within practical given by equation (3): 𝑅𝑆𝑆𝐼[𝑑𝐵𝑚] = [𝑃𝑟 (𝑑)]𝑑𝐵𝑚 = 𝐴 − 10𝑛 log 𝑑𝑖 (3) the measurement distance between target and anchor nodes as shown in equation (4) 𝑑𝑖 = 10 ( 𝐴−𝑅𝑆𝑆𝐼 10𝑛 ) (4) goonjur et al. enhanced the weighted centroid localization algorithm based on … –.| 16 where, 𝑑𝑖 symbolizes each anchor distance to the target node in a meter. a is a strength signal which accepted the distance 1m in dbm. 2.2. centroid localization algorithm (cla) cla is the first and the most straightforward algorithm in range-free. it does not demand the use of any other parameter or rssi to show the distance between a target and anchor nodes. cla in the coordinate information is the only distance information type used, that point out whether a target node is in the anchor node communication range or not. cla is able to communicate with the other node. in addition, it also performances from the theory that each anchor node covers a circular region. to make effective algorithm, the target node utilizes the all anchor nodes location information in its range to measure the centroid position (behnke, r., & timmermann, d., 2008) as follows: 𝑋𝑒𝑠𝑡 = ( 𝑥1+..……+𝑥𝑚 𝑚 ) ; 𝑌𝑒𝑠𝑡 = ( 𝑦1+..……+𝑦𝑚 𝑚 ) (5) (𝑋𝑒𝑠𝑡 , 𝑌𝑒𝑠𝑡 ) shows the target nodes position provide by its two-dimensional coordinates, (𝑥1, 𝑦1) is known as anchor node position, and m is anchor nodes number in the communication range. 2.3. weighted centroid localization algorithm (wcla) the accuracy of cla location estimation has been upgraded by the extent of wcla. wcla introduced anchor node quantification which depends on their distance towards the target nodes. the purpose is to give more effect to the anchor nodes that are closer to the target node. moreover, as the rssi increases with a shorter distance, it is selected as a suitable quantifier. wcla uses weight to guarantee an improved localization compared to the centroid method, where the arithmetic centroid is measured as an object’s location. weight is anchor attraction measurement to object location. the most significant weight value is closer to the target node. to compute the weight, the below formula is used: 𝑤𝑖 = 𝑑𝑖 −𝑔 (6) where 𝑤𝑖 represent the anchor node weight, and g is a level that defines each anchor node contribution. an appropriate value for g was found in one since wcla with a degree of zero is similar to cla. however, the appraisal target’s position is further calculated by the formula: 𝑋𝑒𝑠𝑡 = ∑ (𝑊𝑖×𝑥𝑖) 𝑚 𝑖=1 ∑ 𝑊𝑖 𝑚 𝑖=1 ; 𝑌𝑒𝑠𝑡 = ∑ (𝑊𝑖×𝑦𝑖) 𝑚 𝑖=1 ∑ 𝑊𝑖 𝑚 𝑖=1 (7) the disadvantage of this method cannot entirely reflect the actual condition of the target node due to the interference of path loss and only calculate the target node directly connected with anchor nodes. thus, it largely depends on the anchor nodes density. 2.4. description of the rssi-wcla algorithm an algorithm that combines the rssi and wcla is proposed in this sub-section. based on the free space propagation in an indoor environment (kochláň, m., & miček, j., 2014), the received strength signal can be indicated as: 17 | international journal of informatics information system and computer engineering 1 (2020) 13-22 𝑃𝑅𝑥𝑖 = 𝐴𝑒 𝑆𝑟 = 𝜆2 4𝜋 𝐺𝑅𝑥 . 𝐺𝑇𝑥𝑃𝑇𝑥 4𝜋𝑑2 = 𝑃𝑇𝑥. 𝐺𝑇𝑥 . 𝐺𝑅𝑥 × ( 𝜆 4𝜋𝑑𝑖𝑗 ) 2 (8) the remaining wave power at the receiver from the ith anchor node is 𝑃𝑅𝑥𝑖 , the sender transmission power is 𝑃𝑇𝑥 . 𝐺𝑇𝑥 symbolizes the transmitter antenna gain. 𝐺𝑅𝑥 is the receiver antenna gain. the wavelength is represented by 𝜆. distance between a target node i and anchor node j is represented by 𝑑𝑖𝑗 . the sufficient area of the receiving antenna is symbolized by 𝐴𝑒 and the power density of the signal at the site of the receiving antenna is symbolized by 𝑆𝑟. we can get the relation between 𝑃𝑅𝑥𝑖 and 𝑑𝑖 from equation (8). in the embedded system, the received signal strength is switched into rssi. it is illustrated with the following equation: 𝑅𝑆𝑆𝐼𝑖 = 10𝑙𝑜𝑔 ( 𝑃𝑅𝑥𝑖 𝑃𝑅𝑒𝑓 ) (9) where 𝑃𝑅𝑥𝑖 symbolize the received signal strength from the ith anchor node is, 𝑃𝑅𝑒𝑓 is the reference power. as a result we are able moreover to know the connection between rssi and 𝑃𝑅𝑥𝑖 . from (8), we can modify the equation: 𝑑𝑖𝑗 = 𝜆 × √ 𝑃𝑇𝑥×𝐺𝑇𝑥×𝐺𝑅𝑥 4𝜋𝑃𝑅𝑥𝑖 (10) based on (10), 𝑃𝑅𝑥𝑖 can be expressed as: 𝑃𝑅𝑥𝑖 = 𝑃𝑅𝑒𝑓 × 10 𝑅𝑆𝑆𝐼𝑖 10 (11) substituting formula in (6) to (10), the anchor node weight can be expressed as: 𝑤𝑖 = 𝑑𝑖𝑗 −𝑔 = (𝜆 × √ 𝑃𝑇𝑥×𝐺𝑇𝑥×𝐺𝑅𝑥 4𝜋𝑃𝑅𝑥𝑖 ) −𝑔 = (𝜆 × √ 𝑃𝑇𝑥×𝐺𝑇𝑥×𝐺𝑅𝑥 4𝜋×𝑃𝑅𝑒𝑓×10 𝑅𝑆𝑆𝐼𝑖 10 ) −𝑔 (12) after 𝑤𝑖 normalization, weight can be expressed by: 𝑊𝑖 = 𝑊𝑖 ∑ 𝑊𝑗 𝑚 𝑗=𝑖 = √(10 𝑅𝑆𝑆𝐼𝑖 10 ) 𝑔 ∑ √(10 𝑅𝑆𝑆𝐼𝑗 10 ) 𝑔 𝑚 𝑗=1 (13) thus, the estimation position of the target node can be expressed as below: 𝑋𝑒𝑠𝑡 = ∑ (𝑊𝑖 × 𝑥𝑖 ) 𝑚 𝑖=1 ; 𝑌𝑒𝑠𝑡 = ∑ (𝑊𝑖 × 𝑦𝑖 ) 𝑚 𝑖=1 (14) where, 𝑊𝑖 is as the weight of the anchor node, (𝑥𝑖 , 𝑦𝑖 ) represent the location of the anchor node’s coordinate. from (13) and (14), it can be observed that we are able to get the estimated position of a target node by simply knowing the rssi value and the coordinates of anchor nodes. furthermore, a target node does not need to compute the path loss exponent and obtain other information parameters. additionally, the proposed localization algorithm has the advantage of lower complexity. 3. testbed setup 3.1. results the main focus of the experiments was to deploy within the wsn indoor environment. the real-time experiments were conducted to determine the received strength signal of a wireless sensor network in the testbed environment. the experiment aims to obtain the rssi values between anchor nodes and the target node. the testbed environment is the hall of a laboratory in at our university. the advancement pack utilized within this test is the crossbow telosb sensor node from berkeley university, which uses a cc2420 transceiver with an ieee 802.15.4 standard communication with a built-in 2.4ghz antenna. goonjur et al. enhanced the weighted centroid localization algorithm based on … –.| 18 table 1. indoor positioning algorithms comparison algorithm estimation type indoor precisio n scope los/nlos influence d by multipath cost notes proximity (farid, f., et. al, 2013) signal low to high good both no low a. precision can be made strides by utilizing an extra receiving wire. in any case, it will be expanding the fetched. b. precision in on the arrangement of the estimate within the cells. direction (aoa) (hou, y., et al, 2018) angle of arrival middle good (multipath issue) los yes high a. exactness depends on the antenna’s particular properties. b. the area of the receiving wire must be indicated. time (toa, tdoa) (xu, b., et al., 2013) time of arrival, time difference of arrival high good (multipath issue) los yes high a. needs time to synchronize. b. the area of the radio wire must be indicated. rssi received signal strength indicator high good both no middl e a. require heavy calibration utilized flag engendering show. b. the area of the receiving wire must be indicated. dead reckoning (farid, f., et. al, 2013) velocity low to middle good nlos yes low inaccuracy process is collective. therefore, the deviation in the position settles with the time. map matching (ahmed, a., et al., 2011) projection and pattern recognition middle middle nlos yes middl e a. mapping coordinate simply centers on the calculation and not totally on position strategies, arrange change, and geocoding. b. utilizing design acknowledgment, tall computing complex, and common real-time issues happen. 19 | international journal of informatics information system and computer engineering 1 (2020) 13-22 the testbed consisted of five crossbow telosb sensor nodes programmed with nesc programming language. the program used to converge the data and show the results through the sensor nodes were written in java programming language. for this study, we analyzed the information collected from the nodes which were programmed to send data in real-time. one of the sensor nodes was connected to a laptop which was utilized as the target node, as shown in fig. 1. the remaining four sensor nodes that act as anchor nodes were placed at 0o, 90o, 180o, and 270o distance absence from the target node at the same distance whereas taking the measurements. furthermore, the measurements were taken from a distance of 1m to 5m with an interim of 1m, and at each distance, readings were recorded for 2 minutes, giving a total of 500 readings. the mean value of each node rssi accomplished at a given distance was calculated. fig. 1. illustration position of anchor nodes and target node 4. analysis of experiment results in this section, we analyzed the performance of the rssi-wcla algorithm through the experiment by changing the distance coordinates between the anchor nodes to the target node from 1m to 5m with d1=d2=d3=d4 respectively, and the target node is fixed as the center of measurement. in this experiment, the proposed algorithm is tested for distance 1m, and target node coordinate (1, 1), the anchor nodes coordinate are anchor1 (1, 2), anchor2 (2, 1), anchor3 (1, 0), and anchor4 (0, 1). those anchor nodes obtain the rssi values by the target node, as it is shown in table 2. table 2. the average rssi values of the four sensor nodes distanc e (m) averag e rssi (dbm) for node 1 (0o) averag e rssi (dbm) for node 2 (90o) averag e rssi (dbm) for node 3 (180o) averag e rssi (dbm) for node 4 (270o) 1 -43.3 -45.8 -44.6 -44.2 2 -47.8 -47.2 -47.8 -47.8 3 -48.8 -47.6 -49.4 -49.2 4 -53.4 -55.0 -51.2 -51.4 5 -55.2 -58.0 -53.6 -53.4 the rssi values represent with cycles were spread out at each sensor node according to the location in every room corner, as shown in fig. 2. fig.2 the anchor node location in the room with its rssi goonjur et al. enhanced the weighted centroid localization algorithm based on … –.| 20 1) localization measurement distance error in rssi based on path loss table 3 shows the error distance of rssi according to the path loss measurement between anchor nodes and a target node. from table 1, we can see that increasing the actual distance can increase the error distance; thus, the average error distance shows a 0.88m due to the attenuation of a signal, which is a result of environmental limitations. table 3. measured distance based on path loss real distance (m) average of rssi (dbm) estimation distance (m) error (m) 1 -44.5 1.01 0.01 2 -47.65 1.76 0.24 3 -48.75 2.16 0.84 4 -52.75 5.12 1.12 5 -55.05 7.22 2.22 2) localization measurement distance error based on wcla according to eqn. (7) table 4 shows that the measured distance in wcla is similar to the actual distance, with g = 1. due to wcla only obtaining the information of the anchor node through the position coordinate, we cannot get a significant distance error. table 4. wcla measured distance actual distance (m) center location measured distance (m) 1 1, 1 1 2 2, 2 2 3 3, 3 3 4 4, 4 4 5 5, 5 5 3) localization measurement distance error based on rssi-wcla the algorithm process of rssi-wcla is as follows: step 1: the anchor nodes send their relevant information, including their received signal strength, id information, and position coordinates information to the target node in the form of broadcast. step 2: calculate the weight wi in (13). step 3: calculate the estimated position of the target node in (14) table 5 shows using a similar rssi, and knowing the position of coordinate, we get the average error distance in our proposed algorithm as 0.0552m. the accuracy localization is 1-3% of the actual distance. table 5. rssi-wcla real distance (m) average of rssi (dbm) estimation distance (m) error (m) 1 -44.5 0.992 0.008 2 -47.65 2.056 0.056 3 -48.75 2.904 0.096 4 -52.75 3.936 0.064 5 -55.05 4.948 0.052 in figs. 3 and 4, despite the comparison of the distance between the actual distance and measured distance of the three algorithms under different coordinate numbers, both the anchor nodes relative with the target node, the proposed algorithm shows rssi-wcla obtaining nearly the actual distance value. it is different from the rssi algorithm, where the increase of actual distance led to increasing the error of measured distance significantly. although in fig. 3, it can be observed that the error of distance has a significant influence on the rssi localization performance in wsn when the distance is at a higher level. however, our proposed algorithm shows that the 21 | international journal of informatics information system and computer engineering 1 (2020) 13-22 maximum error distance is less than 0.096m. fig. 3. comparison of actual distance with the result of distance through an experiment 5. conclusion we proposed a localization algorithm that combined rssi and wcla (rssiwcla). comparing this novel algorithm to the rssi algorithm and wcla, our algorithm appears that the exact area of the target hub is more precise, lightweight, and less complicated; due to getting the estimated position of the target node by just mere knowing the rssi values and the coordinates of anchor nodes. the rssi-wcla obtains the average distance error of 0.0552m, whereas the average distance error in rssi is 0.88m, which impacts as it was utilized received strength signal. overall, our results demonstrate a strong effect of indoor localization. the further of our research, we would like to test our algorithms as well as the devices in a large room with more anchor nodes attached to the wall. fig. 4. comparison error of distance on rssi, wlca, and rssi-wlca algorithm references ahmed, a. m., zhu, w., & bekele, t. m. (2011). map-matching and positioning uncertainty in location based services (lbs). in proceedings of the international conference on asia agriculture and animal. behnke, r., & timmermann, d. (2008, march). awcl: adaptive weighted centroid localization as an efficient improvement of coarse grained localization. in 2008 5th workshop on positioning, navigation and communication, 243-250. chen, m., & liu, h. (2012, august). enhance performance of centroid algorithm in wireless sensor networks. in 2012 fourth international conference on computational and information sciences, 1066-1068 goonjur et al. enhanced the weighted centroid localization algorithm based on … –.|22 farid, z., nordin, r., & ismail, m. (2013). recent advances in wireless indoor localization techniques and system. journal of computer networks and communications, 2013. hou, y., yang, x., & abbasi, q. h. (2018). efficient aoa-based wireless indoor localization for hospital outpatients using mobile devices. sensors, 18(11), 3698. kochláň, m., & miček, j. (2014, july). indoor propagation of 2.4 ghz radio signal propagation models and experimental results. in the 10th international conference on digital technologies 2014, 125-129. mukhopadhyay, b., sarangi, s., & kar, s. (2014, february). novel rssi evaluation models for accurate indoor localization with sensor networks. in 2014 twentieth national conference on communications (ncc), 1-6. suzhe, w., & yong, l. (2012, december). node localization algorithm based on rssi in wireless sensor network. in 2012 6th international conference on signal processing and communication systems, 1-4. xu, b., sun, g., yu, r., & yang, z. (2012). high-accuracy tdoa-based localization without time synchronization. ieee transactions on parallel and distributed systems, 24(8), 1567-1576. zhang, s., & xing, t. (2013, december). open wsn indoor localization platform design. in 2013 2nd international symposium on instrumentation and measurement, sensor network and automation (imsna), 845-848. 1 | international journal of informatics information system and computer engineering 1 (2020) 35-46 air quality prediction in smart city’s information system ivan kristianto singgih department of the industrial and systems engineering, korea advanced institute of science and technology, daejeon, 305-701, republic of korea correspondence: e-mail: ivanksinggih@gmail.com a b s t r a c t s a r t i c l e i n f o the introduction of new technology and computational power enables more data usages in a city. such a city is called a smart city that records more data related to daily life activities and analyzes them to provide better services. such data acquisition and analysis must be conducted quickly to support real-time information sharing and support other decision-making processes. among such services, an information system is used to predict the air quality to ensure people's health in the city. the objective of this study is to compare various machine learning techniques (e.g., random forest, decision tree, neural network, naïve bayes, etc.) when predicting the air quality in a city. for the comparison, we perform the removal of records with empty values, data division into training and testing datasets, and application of the k-fold cross-validation method. numerical experiments are performed using a given online dataset. the results show that the three best methods are random forest, gradient boosting, and knearest neighbors with precision, recall, and f1-score values more than 0.63. article history: received 8 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 vailable online 09 sep 2018 ___________________ _ keywords: definitions of shophouses; identity; george town; influence architecture and design. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 1 (2020) 35-46 i k singgih. air quality prediction in smart city’s information system | 36 1. introduction mart city integrates the physical world and the virtual world. a concept used for performing such connectivity is called digital twin that is a virtual model representing the physical world (marr, 2020). using such a virtual model allows monitoring the physical system, preventing problems from happening, finding new opportunities, and planning the future. the interaction is illustrated in smart city korea (2020) through fig. 1. massive iots, digital twins, and data hubs are utilized to generate the required information in the integration process. in the smart city, fixed/mobile sensors are installed within the city to observe real behaviors (e.g., the people) and conduct a better operation of the virtual world. there are various subareas within the smart city, including smart mobility, smart buildings, etc. among them, the smart environment is the one that manages air pollution control to ensure the health of the citizens (alvear et al., 2018). the smart environment system's continuous improvement is supported by the wide use of the internet of things that provides better connectivity between multiple sensors located in the dispersed area and ease the air quality monitoring process (zhang & woo, 2020). the existence of different air pollutants causes harm to the respiratory systems. such air pollutants are nitrogen dioxide (no2), carbon monoxide (co), ozone (o3), sulphur dioxide (so2), and particulate matter (pm). real-time monitoring stations are built by many cities to check the air quality, then inform people when it is safe to conduct outside activities and plan better movements (zhang & woo, 2020). systems for collecting and assessing air quality have been installed in several areas, e.g., peking university (with 100 thousand data from 30 devices) (hu et al., 2019), christchurch that is a part of ibm’s smart city initiatives (marek et al., 2017), los angeles (wu et al., 2017), etc. various information systems are implemented for supporting the data collection and air quality information transfer to the people. an example of the information system used for air quality management in los angeles is presented in fig. 2 (wu et al., 2017). in this implementation, a remote data collection can be performed using a smartphone. the collected image data are analyzed using machine learning to calculate the particle concentration in the air and evaluate the air quality. 37 | international journal of informatics information system and computer engineering 1 (2020) 35-46 fig. 1. interaction between physical and virtual world in smart city concept i k singgih. air quality prediction in smart city’s information system | 38 fig. 2. information system for air quality management in los angeles 2. methodology information system in smart cities has a component for the data acquisition and a server to store and process the obtained data. the good adoption of technologies for data collection and computation determines the success of smart city developments (marek et al., 2017). given that multiple sources of data (e.g., open data, online data sharing) emerge, improving the information system interoperability, including how to utilize existing data, is a great challenge to be solved in smart city projects. many smart city initiatives have been started. one of them is era-planet, a wide european network that consists of 118 researchers in 35 institutions and 18 countries (tsinganos et al., 2017). the architecture of the information system related to the air quality sensing process performs the following tasks (alvear et al., 2018): 1) sampling the sampling task measures the pollutant in the air that includes the calibration process. by performing such sampling with many mobile sensors, the problem of sampling error can be handled because of the possibility of considering redundant data and statistical analysis. 39 | international journal of informatics information system and computer engineering 1 (2020) 35-46 2) data filtering through the filtering process, redundant data and wrong measurements are removed. 3) data transfer the collected data are uploaded from the sensors to the cloud (servers). the upload process is managed based on some iot protocols. 4) data processing the observed data are processed to obtain a conclusion on the air quality. through this process, a pollution distribution map is generated. 5) presentation of the analysis result the results can be presented as a graphical map. the architecture itself can be divided into three layers (schürholz et al., 2020): 1) data layer the data layer contains a database of historical data and prediction data. 2) logic layer the logic layer converts the input data before being used in the analysis and performs the prediction process. 3) visualization layer the visualization layer passes the information to be visualized in the enduser devices. the introduction of inexpensive small sensors allows retrieving a huge amount of data in real-time fashion (hu et al., 2019). effective machine learning techniques are implemented in this study to perform such a real-time air quality assessment. the machine learning techniques used in this study are listed in table i. the selected methods have been proven to perform well for predicting air quality purposes. studies that used each method (or its variants) are presented as well. 3. results and discussion we use the python sklearn library (pedregosa et al., 2011) to implement the machine learning techniques. the code is written using the visual studio 2019 community platform. a partial view of the code is presented in fig. 3. air quality prediction data presented in (bhat, 2020) is solved. among 26,6219 data, we remove records with any empty values and obtain 4,646 records to be used in our study. we exclude the location and time stamp fields from the observed data. the preprocessed data are stored in an excel input file and is imported into python. libraries for performing the calculations and generating a graphical representation of the results are used. the dependent variable is the air quality with the following values: severe, very poor, poor, moderate, satisfactory, good. the independent variables are: 1) pm2.5 pm is the abbreviation of particulate matter that includes potential harmful compounds, which can reach human respiratory systems (chaparro et al., 2020). pm2.5 refers to cases of air i k singgih. air quality prediction in smart city’s information system | 40 particles with the mass per cubic meter less than 2.5 µm. 2) pm10 3) no no refers to nitrogen oxide. 4) no2 no2 refers to nitrogen dioxide. 5) nox nox is the total amount of no and no2. 6) nh3 nh3 refers to ammonia. 7) co co refers to carbon monoxide. 8) so2 so2 refers to sulfur dioxide. 9) o3 o3 refers to ozone. 10) benzene 11) toluene 12) xylene analysis steps performed in this study are: 1) dividing the dataset into training and testing data in our implementation, the percentage of testing data is set into 20%. the data is shuffled before the division. 2) testing the accuracy of each technique using the k-fold crossvalidation the number of used splits is 10. the training data is shuffled as well before performing the testing. 3) fitting the testing data table 1. used machine learning techniques in this study machine learning technique reference adaptive boosting (ab) liu and chen (2020) linear classifiers with stochastic gradient descent training (sgd) ganesh et al. (2017) neural network (multi-layer perceptron a) (nnmlp) ganesh et al. (2017), gu et al. (2020), sun et al. (2020), wang et al. (2017), zhao et al. (2020) gradient boosting (gb) zhang et al. (2019), zhang et al. (2019b), liu et al. (2019), yu et al. (2016), feng et al. (2018) random forest (rf) liu et al. (2019), yu et al. (2016), feng et al. (2018) k-nearest neighbors (knn) zhao et al. (2020) decision tree (cart) zhang et al. (2019b) naive bayes (gaussian a) (nb) feng et al. (2018), melgarejo et al. (2015) support vector machine (c-support vector a) (svm) ganesh et al. (2017), gu et al. (2020), liu et al. (2019), melgarejo et al. (2015), dun et al. (2020) 41 | international journal of informatics information system and computer engineering 1 (2020) 35-46 aspecific one considered in this study. fig. 3. a partial view of the code fig. 4. true positive, false positive, false negative, and true negative cases precision, recall, f1-score, and support metrics are measured for each technique. true positive, false positive, false negative, and true negative cases are observed to calculate such values. the cases are defined in fig. 4 based on the comparison between results concluded by the test and the real data (parikh et al., 2008). definition and formula of precision, recall, f1-score, and support are presented in table ii. machine learning methods that can predict the air quality better are the ones with higher scores. result of accuracy testing using kfold cross validation is presented using boxplots in fig. 5. three techniques that have the best accuracy are rf, gb, and knn. these three methods have good i k singgih. air quality prediction in smart city’s information system | 42 average accuracy and a smaller deviation in the accuracy calculation when considering different training and testing datasets, compared with the others. the worst accuracies are obtained by sgd and svm methods. table 2. definition and formula of precision, recall, f1-score, and support metric definition formula precision total number of retrieved data that are relevant/total number of retrieved data (ting, 2011) tp / (tp + fp) (jiang et al., 2017) recall total number of retrieved data that are relevant/total number of relevant data in the database (ting, 2011) tp / (tp + fn) (jiang et al., 2017) f1-score a weighted value obtained from the precision and recall values with 1 as its best value and 0 as its worst 2 * precision * recall / (precision + recall) (yuan et al., 2020) support number of occurrences of each class in y_true fig. 5. accuracy comparison of the machine learning techniques table 3. average value of each metric using the testing data machine learning technique prediction recall f1score ab 0.36 0.52 0.41 sgd 0.38 0.41 0.39 nnmlp 0.73 0.64 0.66 gb 0.82 0.77 0.79 rf 0.81 0.76 0.78 knn 0.81 0.77 0.79 cart 0.75 0.71 0.73 nb 0.65 0.70 0.67 svm 0.22 0.19 0.14 table 4. detailed metric values of rf method class predicti on reca ll f1scor e suppo rt good 0.84 0.64 0.73 67 satisfact ory 0.78 0.84 0.81 269 moderat e 0.83 0.86 0.84 383 poor 0.75 0.69 0.72 110 very poor 0.81 0.81 0.81 77 severe 0.85 0.71 0.77 24 43 | international journal of informatics information system and computer engineering 1 (2020) 35-46 table 5. detailed metric values of gb method class predicti on reca ll f1scor e suppo rt good 0.85 0.66 0.74 67 satisfact ory 0.80 0.84 0.82 269 moderat e 0.84 0.86 0.85 383 poor 0.72 0.68 0.70 110 very poor 0.81 0.79 0.80 77 severe 0.90 0.79 0.84 24 table 6. detailed metric values of knn method class predicti on reca ll f1scor e suppo rt good 0.70 0.63 0.66 67 satisfact ory 0.76 0.81 0.78 269 moderat e 0.84 0.84 0.84 383 poor 0.77 0.73 0.75 110 very poor 0.86 0.81 0.83 77 class predicti on reca ll f1scor e suppo rt severe 0.90 0.79 0.84 24 the fitting results of the testing data are presented in tables iii-vi. in table iii, the average value of each metric calculated from all classification class is presented. the detailed metric values for the three best techniques are presented in tables iv-vi. in these tables, evaluation is performed for each class (severe, very poor, poor, moderate, satisfactory, good). it can be seen that the value of each metric is similar for each class when a certain method is implemented. 4. conclusion in this study, we implement several machine learning techniques to predict air quality as part of the smart city's information system. based on the numerical experiments, random forest, gradient boosting, and k-nearest neighbors have the best accuracies. future studies must assess whether it is necessary to include all input values in the models and consider how to deal with incomplete records. references alvear, o., calafate, c. t., cano, j.-c., & manzoni, p. (2018). crowdsensing in smart cities: overview, platforms, and environment sensing issues. sensors, 18(2), 460. bhat, n. (2020, october 1). air quality level of different cities in india (2015-2020). kaggle dataset. https://www.kaggle.com/nareshbhat/air-quality-pre-and-postcovid19-pandemic i k singgih. air quality prediction in smart city’s information system | 44 chaparro, m. a. e., chaparro, m. a. e., castañeda-miranda, a. g., marié, d. c., gargiulo, j. d., lavornia, j. m., natal, m., & böhnel, h. n. (2020). fine air pollution particles trapped by street tree barks: in situ magnetic biomonitoring. environmental pollution, 266(1), article 115229. dun, m., xu, z., chen, y., & wu, l. (2020). short-term air quality prediction based on fractional grey linear regression and support vector machine. mathematical problems in engineering, 2020(1), article 8914501. feng, c., tian, y., gong, x., que, x., & wang, w. (2018). mcs-rf: mobile crowdsensing–based air quality estimation with random forest. international journal of distributed sensor networks, 14(10), 1–15. ganesh, s. s., arulmozhivarman, p., & tatavarti, r. (2017). forecasting air quality index using an ensemble of artificial neural networks and regression models. journal of intelligent systems, 28(5), 893–903. gu, k., zhou, y., sun, h., zhao, l., & liu, s. (2020). prediction of air quality in shenzhen based on neural network algorithm. neural computing and applications, 32(7), 1879–1892. hu, z., bai, z., bian, k., wang, t., & song, l. (2019). real-time fine-grained air quality sensing networks in smart city: design, implementation, and optimization. ieee internet of things journal, 6(5), 7526–7542. jiang, c., liu, y., ding, y., liang, k., & duan, r. (2017). capturing helpful reviews from social media for product quality improvement: a multi-class classification approach. international journal of production research, 55(12), 3528–3541. liu, h., & chen, c. (2020). spatial air quality index prediction model based on decomposition, adaptive boosting, and three-stage feature selection: a case study in china. journal of cleaner production, 265(1), article 121777. liu, h., li, q., yu, d., & gu, y. (2019). air quality index and air pollutant concentration prediction based on machine learning algorithms. applied sciences, 9(19), 4069. marek, l., campbell, m., & bui, l. (2017). shaking for innovation: the (re)building of a (smart) city in a post disaster environment. cities, 63(1), 41–50. marr, b. (2020, october 3). what is digital twin technology and why is it so important? forbes. https://www.forbes.com/sites/bernardmarr/2017/03/06/what-isdigital-twin-technology-and-why-is-it-so-important/#3388188e2e2a melgarejo, m., parra, c., & obregón, n. (2015). applying computational intelligence to the classification of pollution events. ieee latin america transactions, 13(7), 2071–2077. 45 | international journal of informatics information system and computer engineering 1 (2020) 35-46 parikh, r., mathai, a., parikh, s., sekhar, g. c., & thomas, r. (2008). understanding and using sensitivity, specificity and predictive values. indian journal of ophthalmology, 56(1), 45–50. pedregosa, f., varoquaux, g., gramfort, a., michel, v., thirion, b., grisel, o., blondel, m., prettenhofer, p., weiss, r., dubourg, v., vanderplas, j., passos, a., cournapeau, d., brucher, m., perrot, m., & duchesnay, é. (2011). scikit-learn: machine learning in python. journal of machine learning research, 12(1), 2825–2830. schürholz, d., kubler, s., & zaslavsky, a. (2020). artificial intelligence-enabled context-aware air quality prediction for smart cities. journal of cleaner production, 271(1), 121941. smart city korea. (2020, october 3). introducing core 1, 2, and 3 smart city innovation growth engines. ministry of land, infrastructure and transport. https://smartcity.go.kr/en/ sun, x., xu, w., jiang, h., & wang, q. (2020). a deep multitask learning approach for air quality prediction. annals of operations research. ting, k. m. (2011). precision and recall. in encyclopedia of machine learning, 781-781. springer, boston, ma. tsinganos, k., gerasopoulos, e., keramitsoglou, i., pirrone, n., & the era-planet team. (2017). era-planet, a european network for observing our changing planet. sustainability, 9(6), 1040. wang, j., zhang, x., guo, z., & lu, h. (2017). developing an early-warning system for air quality prediction and assessment of cities in china. expert systems with applications, 84(1), 102–116. wu, y.-c., shiledar, a., li, y.-c., wong, j., feng, s., chen, x., chen, c., jin, k., janamian, s., yang, z., ballard, z. c., göröcs, z., feizi, a., & ozcan, a. (2017). air quality monitoring using mobile microscopy and machine learning. light: science & applications, 6(1), 17046. yu, r., yang, y., yang, l., han, g., & move, o. a. (2016). raq–a random forest approach for predicting air quality in urban sensing systems. sensors, 16(1), article 86. https://doi.org/10.3390/s16010086 yuan, j., zhang, l., guo, s., xiao, y., & li, z. (2020). image captioning with a joint attention mechanism by visual concept samples. acm transactions on multimedia computing, communications, and applications, 16(3), article 83. https://doi.org/10.1145/3394955 zhang, d., & woo, s. s. (2020). real time localized air quality monitoring and prediction through mobile and fixed iot sensing network. ieee access, 8(1), 89584– 89594. https://doi.org/10.1109/access.2020.2993547 i k singgih. air quality prediction in smart city’s information system | 46 zhang, y., wang, y., gao, m., ma, q., zhao, j., zhang, r., wang, q., & huang, l. (2019a). a predictive data feature exploration-based air quality prediction approach. ieee access, 7(1), 30732–30743. https://doi.org/10.1109/access.2019.2897754 zhang, y., zhang, r., ma, q., wang, y., wang, q., huang, z., & huang, l. (2019b). a feature selection and multi-model fusion-based approach of predicting air quality. isa transactions, 100(1), 210–220. https://doi.org/10.1016/j.isatra.2019.11.023 zhao, z., qin, j., he, z., li, h., yang, y., & zhang, r. (2020). combining forward with recurrent neural networks for hourly air quality prediction in northwest of china. environmental science and pollution research, 27(23), 28931–28948. https://doi.org/10.1007/s11356-020-08948-1 zhao, x., song, m., liu, a., wang, y., wang, t., & cao, j. (2020). data-driven temporalspatial model for the prediction of aqi in nanjing. journal of artificial intelligence and soft computing research, 10(4), 255–270. https://doi.org/10.2478/jaiscr-20200017 1 | international journal of informatics information system and computer engineering 1 (2020) 23-34 computer & network equipment management system (cnemas) application measurement imelda pangaribuan*, a rahman, s mauluddin information system department, universitas komputer indonesia, jl. dipatiukur no: 112 116, bandung, indonesia correspondence: e-mail: *imelda@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o pt. abc is one of the state-owned enterprises in indonesia. pt. abc has policy to give some equipment for employees to rent. application use to control all equipment that rent by employee named as cnemas (computer & network equipment management system nasional). there are some problems faced by pt. abc in using cnemas, such as accuracy and its roles at pt. abc business process. objective of this research is to measure maturity of cnemas. method used in this research are structured approach method as analysis method and information technology assurance framework (itaf) used as measurement method. tools such as flowcharts, context diagram and data flow diagram used to describe the process in cnemas. the result of this research are gap found at ac2 (awareness and communication component) where the gap is 0.71. the conclusion is cnemas needs improvement in accuracy, equipment location, on time report production, and consistency. article history: received 6 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 09 sep 2018 __________________ keywords: cnemas, computer network, application measurement, itaf international journal of informatics, information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 1 (2020) 23-34 pangaribuan, et al. computer & network equipment management system (cnemas)...|24 i. introduction pt. abc is a state-owned enterprise (bumn) which is engaged in information and communication technology (ict) services and telecommunications networks in indonesia. in an effort to transform into a digital telecommunication company, pt. abc group certainly has divisions to implement business strategies and operations of customer-oriented companies. one division that affects the business process of pt. abc is an it division that has responsibility as an information provider, and as a center for the development of information technology (it). in the it division at pt. abc is broken down again into sub-divisions that have different responsibilities, one of which is a small sub-division namely digital work place (dwp) at pt. abc witel (persero) bandung lembong branch. dwp itself has responsibility in providing equipment both computer and non-computer ordered from pt.pins (a subsidiary of pt. abc) which will later be used by employees of pt abc. in the internal business processes that exist in the dwp currently supported by it in the form of computer infrastructure, servers, and networks, application systems, and databases that refer to the center of pt. abc in jakarta. the dwp section aims to control all equipment used so that when the contract period of the equipment is discharged the equipment must be appropriate. there are some research had done to audit applications. (z. rezaee et al., 2001) has observed that by the time the e-commerce technologies and internet were adopted by organizations, there had been some changes in business practices and the process of storing and processing business transactions. therefore, the auditors need to perform audits continuously, to ensure that the applications performed as expected and to conduct internal controls. (abu musa et al., 2004) stated that the audit of ebusiness had been a challenge for external auditors. then, (junaid m. shaikh et al., 2005) also discussed that by the time an organization's business relies on information technology, the auditors should apply electronic auditing (ea) framework associated with the technologies adopted. ea implementation is part of the computerassisted auditing techniques (caats). cnemas (computer & network equipment management system nasional) is one of the applications to support the dwp part work process in controlling equipment that is leased to employees of pt. abc indonesia in order to achieve success and goals for the construction of pt. abc indonesia. based on the results of the interview regarding cnemas, there are some obstacles in the work process that are in the dwp. problem face are pt. abc cannot reach 100% equipment return, monitoring cannot be updated automatically, the dwp does not know yet maturity level of the application and the gap and human error in the use of equipment unidentified yet. the cause of the problem is not yet known in detail by the dwp, thus inhibiting the purpose of making cnemas. so, the dwp is still experiencing problems in information technology governance and making appropriate policy recommendations in the future development of the cnemas application. because this will greatly affect dwp business process performance. based on these problems, cnemas application needs to be measured. as the solution offered here 25 | international journal of informatics information system and computer engineering 1 (2020) 23-34 is audit cnemas using itaf which is one method that can be used to audit applications effectively. in this study, it only focuses on how cnemas controls the devices that are leased to employees of pt. abc as well as knowing the maturity level and the cnemas gap which later will be seen whether cnemas is in line with what has not been expected with pt. abc. research method used is a descriptive approach, then the data collection method used is observation, interview and questionnaire. then the system approach method used is a structured approach. the purpose of the research is to know the maturity level and the cnemas gap, which later would show whether cnemas was in accordance with what expected by pt. abc. 2. literature review there are several methods are used in this research. information technology assurance framework (itaf) is used as framework to measure maturity level of cnemas application. structures approach was used to analyze the application. 2.1 itaf method used in this research is information technology assurance framework (itaf). itaf is a comprehensive and good-practicesetting reference model that: 1. establishes standards that address is audit and assurance professional roles and responsibilities; knowledge and skills; and diligence, conduct and reporting requirements. 2. defines terms and concepts specific to is assurance. 3. provides guidance and tools and techniques on the planning, design, conduct and reporting of is audit and assurance assignments 2. 2 structured approach structured systems analysis and design method (ssadm), originally released as methodology, is a systems approach to the analysis and design of information systems. ssadm was produced for the central computer and telecommunications agency, a uk government office concerned with the use of technology in government, from 1980 onwards (mike et al., 1999). the structured development approach includes structured analysis and structured design. structured analysis focuses on identifying facts of the system and limitations of the system in order to find out new system requirements. it focuses on what are the features and facilities that the new system requirements (sajja et al., 2017). 2.3 information technology maturity level information technology maturity refers to an organization’s capability to utilize its existing it infrastructure to obtain business value (lin et al., 2013). maturity assessment involves scoring the organization against defined criteria and a ranking scheme. this assessment is generally organized in ascending steps with strategies on how to move up the maturity scale. scales are often defined in a 1–5 range that indicates increasing levels of maturity (lin et al., 2013; curley et al., 2016). according to nunnally & bernstein measurement can be defined as a process https://en.wikipedia.org/wiki/central_computer_and_telecommunications_agency https://en.wikipedia.org/wiki/central_computer_and_telecommunications_agency https://en.wikipedia.org/wiki/uk_government https://en.wikipedia.org/wiki/uk_government pangaribuan, et al. computer & network equipment management system (cnemas)...|26 of giving numbers or labels to attributes with standardized rules or agreed upon to represent measured attributes (nunnally et al., 1994). according to mardapi measurement is basically the activity of determining the number of an object systematically (mardapi et al., 2004). according to djemari mardapi assessment is the activity of interpreting or describing the results of measurements (mardapi et al., 2004). according to the isaca literature module measurement is a value both large and small to assess the maturity of an object under study and for the standard it has been determined for each existing domain, to measure results is done by adding the score of each respondent's answer later. will be divided based on the number of respondents (isaca et al., 1969). according to the isaca literature module the level of maturity is a process of assisting in defining understanding, for the maturity level there are 6 levels of maturity, level 0 which means no process, the second is level 1 where it process has done but has no procedure, the third is level 2 where organization already done the it process but it hasn't been implemented yet, then the fourth is level 3 where organization has already done the it process and the implementation is still there, then the fifth is level 4 where the it process has been done and already has the procedure and implementation fixed, and the last level 5 is perfect (isaca et al., 1969) 3. results the research method is a series of activities that provide an overview of the steps in conducting a particular research. the research method used by the current writer is the method of descriptive analysis giving a systematic description of the collection of facts on an object of research using certain data collection methods. the method of gathering data that i use is by conducting observation interviews and questionnaires as a method of collecting primary data and documentation as one of the secondary data collection methods. the ecosystem approach method used by the authors in this study is a structured approach using analytical tools such as flowcharts. research objective the purpose of the research design is to know the maturity level and the cnemas gap, which later will show whether cnemas is in line with what has been desired by pt. abc. overview of the running system running system describe using flowchart that shown by fig. 1 and data flow inside the system shows at data flow diagram at fig. 2. 27 | international journal of informatics information system and computer engineering 1 (2020) 23-34 fig. 1. flowchart of running system pangaribuan, et al. computer & network equipment management system (cnemas)...|28 fig. 2. data flow diagram procedure of cnemas application shown at fig. 1 in flowchart form and the steps are describe below: 1. user sending employee data to human resource department (hrd) 2. hrd entering data to cnemas application. 3. dwp then create equipment request data to pt. pins 4. pt. pins sending equipment data to dwp, then dwp entering equipment data to cnemas. 5. it division create report base on employee data and equipment data and distribute it to egm dwp . measurement that had done focus on the process of planning and aligning cnemas application with company objectives, including control issues, tactics and identification of how cnemas can contribute maximally to achievement objectives of cnemas application. base on the aim of research application control (ac) domain is the correct domain to measure cnemas application. below are sub of application control (ac) domain: 1. ac1 – preparation and authorization of data sources 2. ac2 – collection and entry of data sources 29 | international journal of informatics information system and computer engineering 1 (2020) 23-34 3. ac3 – examination of accuracy, completeness and authenticity 4. ac4 – integrity and validation of inputted data 5. ac5 – output, reconciliation and handling problems the reason of using application control (ac) domain is because ac domain discusses how cnemas start to prepare data source authorization, collecting input data, accuracy, completeness and authenticity of the data, validating data and output as well as handling errors that occur at cnemas and other reasons namely in other domains. suitable and not carried out at cnemas application. questionnaire and data collection determination of value and maturity level using questionnaire distributed to respondents who have been determined based on responsibilities, where for parties from it department act as it managers and act as users of the cnemas system so that they have the ability to provide assessments related to implementation. the questionnaire was designed in the format of column that can be filled with weighted values that have been set. the questions are grouped according to ac1 ac6, and in each group the questions will involve 2 columns which will represent the current maturity value (as is) and expected maturity level (to be). each question has 6 answer choices that indicate the level of maturity of the maturity attribute in each itaf process. respondents can choose one of the answers that is considered to best represent the conditions of maturity both current and expected. the weight of the answer is adjusted to the current level of cnemas application maturity (maturity value) and expected (maturity target) shown at table 1: table 1. score weight of answer no maturity score weight 1 a 0 2 b 1 3 c 2 4 d 3 5 e 4 6 f 5 score of the calculated value of the maturity attribute will be rounded to produce the maturity level of cnemas. itaf level starts from 0 (none) to 5 (optimal). below are meaning of maturity score at itaf: a = 0, which means non-existent (there is no process in the application). b = 1, it means initial / ad hoc (done, but there is no procedure yet). c = 2, meaning repeatable but intuitive (done, but not fixed at the moment). d = 3, it means defined process (done and fixed). e = 4, meaning managed and measureable (there are procedures, and there are standards and there is monitoring). f = 5, meaning optimized (perfect, the application runs well and the company quickly adapts to changes). the score of answer questionnaire showing maturity index and shows the maturity level. maturity index and maturity level shows by tables 2 and 3. pangaribuan, et al. computer & network equipment management system (cnemas)...|30 level maturity index maturity level 0 0 – 0.49 non existense (there is no it process) 1 0.50 – 1.49 initial ad hoc (has it process but does not have procedure) 2 1.50 – 2.49 repeatable but intuitive (has it process and procedure but inconsistent) 3 2.50 – 3.49 defined process (has done it process and procedure and consistent) 4 3.50 – 4.49 managed and measurable (implement it process, procedure and monitoring) 5 4.50 – 5 optimized (perfect, it fully implemented and organization easy to adapt with change) 1) maturity level questionnaire data results table 3. accumulation of questionnaire answers domain maturity score maturity target gap ac 1 28 30 2 ac 2 30 35 5 ac 3 25 25 0 ac 4 50 50 0 ac 5 25 25 0 ac 6 15 15 0 numbers on table 3 are obtained based on the domain of maturity value and maturity target filled by the four respondents then accumulated using excel formula in general where: 1) attribute maturity value = (total answer maturity value of four respondents (based on each domain) / number of respondents. 2) attribute maturity targets = (total answers of the four maturity targets respondents) / number of respondents. then use excel formula like below to get the accumulated value. table 2. maturity level score 31 | international journal of informatics information system and computer engineering 1 (2020) 23-34 =sum (ac 1 r1+ ac 2 r2+ ac 3 r3 + ac 4 r4) / respondents number. 2) maturity level based on table 4, it can be said that the current maturity level in the ac 1 process is related to the cnemas application at pt. abc, which is overall in level 5 (optimized) with the conditions that have been achieved by the tools function, has run smoothly as expected, the document components have been arranged according to existing standards, the unit has been arranged neatly along with unit identification in accordance with the expectations of pt. abc, and the return of documents that have not been validated have been arranged according to existing standards while the expected level of maturity is at level 5 (optimized) where the system process has been optimized, runs well and quickly adapts to changes but there are some parts that must be corrected. from the comparison of the current maturity level with the expected one, the gap value is 0.33. this shows that cnemas almost reached the expectations desired by pt. abc and needed recommendations for improvements from the maturity level gap. table 4. maturity level ac 1 – ac 6 for ac 2 it can be said that the current maturity level in the ac 2 process is related to the cnemas application at pt. abc is in total located in 4 managed and measureable (conducted there is a procedure, and standard and monitoring) which in this domain cnemas has determined who can access cnemas in accordance with expectations, the wrong procedure for document repairs is in line with expectations of pt. abc, if there is a data transaction or an error occurs on a failed report, it can be immediately addressed and the document is stored safely in accordance with the expectations of pt. abc, while the expected level of domain number of questions maturity score maturity target gap ac 1 6 28 30 2 index average 4,67 5 0,33 ac 2 7 30 35 5 index average 4,29 5 0,71 ac 3 5 25 25 0 index average 5 5 0 ac 4 10 50 50 0 index average 5 5 0 ac 5 6 25 25 0 index average 4,17 4,17 0 ac 6 3 15 15 0 index average 5 5 0 pangaribuan, et al. computer & network equipment management system (cnemas)...|32 maturity is at level 5 (optimized) where the system process has been optimized, runs well and quickly adapts to changes. from the comparison of the current maturity level with the expected one, the gap value is 0.71. this shows that the achievement of the maturity level target expected by pt abc to cnemas and needs improvement in terms of repairs to transactions and failed reports can be resolved quickly so that it does not take much longer. ac3, it can be said that the current maturity level in the ac 3 process is related to the cnemas application at pt. abc is on the whole level 5 (optimized) while the expected level of maturity is level 5 (optimized) where the system process has been optimized, runs well and quickly adapts to changes. from the comparison of the current maturity level with the expected one, it shows a gap value of 0. this shows that cnemas has achieved the desired expectations by pt. abc and does not need recommendations for improvement from the maturity level gap but must maintain consistency. ac 4, it can be said that the current maturity level in the ac 4 process is related to the cnemas application at pt. abc is on the whole level 5 (optimized) while the expected level of maturity is level 5 (optimized) where the system process has been optimized, runs well and quickly adapts to changes. from the comparison of the current maturity level with the expected one, it shows a gap value of 0. this shows that cnemas has achieved the desired expectations by pt. abc and does not need recommendations for improvement from the maturity level gap but must maintain consistency. it can be said that the current maturity level in the ac 5 process is related to the cnemas application at pt. abc as a whole is in 4 managed and measurable (performed there is a procedure, and is standard and there is monitoring) while the expected level of maturity is at the level of 4 managed and measurable (conducted procedure, and standard and monitoring). from the comparison of the current maturity level to the expected one shows a gap value of 0. this shows that the achievement of the maturity level target expected by pt abc to cnemas and does not require recommendations for improvement from the maturity level gap. it can be said that the current maturity level in the ac 6 process is related to the cnemas application at pt. abc is on the whole level 5 (optimized) while the expected level of maturity is level 5 (optimized) where the system process has been optimized, runs well and quickly adapts to changes. from the comparison of the current maturity level with the expected one, it shows a gap value of 0. this shows that cnemas has achieved the desired expectations by pt. abc and does not need recommendations for improvement from the maturity level gap but must maintain consistency. 3) recommendation table 3 showing recommendation base on measurement result of cnemas application: 33 | international journal of informatics information system and computer engineering 1 (2020) 23-34 tabel 5. recommendation 4. conclusion lorem based on the results of the questionnaire pt. abc can find out the value of maturity and maturity targets of cnemas but there are still a few problems in some of its domains in cnemas's performance which are the scope of research. these problems include the following: 1. the current maturity level in the ac 1 process is related to the cnemas application at pt. abc, which is overall in level 5 (optimized), but there are some achievements that are not in line with expectations while the expected level of maturity is at level 5 (optimized) where the system process has been optimized, runs well and quickly adapts to changes. from the comparison of the current maturity level with the expected one, the gap value is 0.33. this shows that cnemas almost reached the expectations desired by pt. abc and needed recommendations for improvements from the maturity level gap. 2. the current maturity level in the ac 2 process is related to the cnemas application at pt. abc as a whole is in 4 managed and measureable (performed there is a procedure, and standard and there is monitoring) but there are some achievements that are not in line with expectations while the expected level of maturity is at level 5 (optimized) where the system process is optimized and quickly adapt to change. from the comparison of the current maturity level with the expected one, the gap value is 0.71. this shows that the maturity level target that pt abc has not yet achieved in cnemas and needs recommendations for improvement is the maturity level gap. 5. acknowledgment publication of this research support by rector of universitas komputer indonesia and journal division of computer indonesia. references abu-musa, a. a. (2004). auditing e-business: new challenges for external auditors. journal of american academy of business, 4(1), 28-41. capability maturity model integration (cmmi) version 1.2 overview,http://www.sei.cmu.edu/library/assets/cmmi-overview071.pdf domain maturity level now maturity level hope recommendation ac 4 managed and measureabel 5 optimized overall, cnemas almost reached expectations, but pt. abc still has to improve the performance of cnemas on the accuracy of device location and the timeliness in making reports to be more effective and must maintain consistency so that the performance of cnemas is not far from the expectations desired by pt. abc. pangaribuan, et al. computer & network equipment management system (cnemas)...|34 software engineering institute, carnegie mellon university, 2007. access time november 21 2018 curley, m., kenneally, j., & carcary, m. (2016). it capability maturity frameworktm (it-cmftm). van haren. isaca, “modul literatur isaca”, 1969. lin, c., huang, y. a., li, c. f., jalleh, g., & liu, y. c. (2013). a preliminary study of key factors affecting management and evaluation of it outsourcing contracts in hospitals. in handbook of research on icts and management systems for improving efficiency in healthcare and social care (pp. 1109-1129). igi global. mardapi , “definisi pengukuran”, 1st ed, yogyakarta : gava media, 2004. mike, g., & karel, r. (1999). history of ssadm, ssadm an introduction. nunnally, j. c. (1994). psychometric theory 3e. tata mcgraw-hill education. rezaee, z., elam, r., & sharbatoghlie, a. (2001). continuous auditing: the audit of the future. managerial auditing journal. sajja, p. s. (2017). structured systems development approach. in essence of systems analysis and design. springer, singapore. shaikh, j. m. (2005). e‐commerce impact: emerging technology–electronic auditing. managerial auditing journal. 20(4), 408 – 421. 77 | international journal of informatics information system and computer engineering 2(1) (2021) 77-82 virtual voting system gaurav kumar*, smriti gupta**, divya agarwal***, astha tiwari**** department of information technology, noida institute of engineering and technology, greater noida, india e-mail: *yaduyash@niet.co.in a b s t r a c t s a r t i c l e i n f o india's voting system plays an important role in indian democracy. the existing system is offline and has certain weaknesses. in recent years, the spread of covid19, inefficient rural voters, people far from their place of birth, paper waste affecting nature, budgets that should be used for development, invisible fraud, waste of human labor, have been recorded and can be avoided by the virtual voting system. the research aims to supply an easy and secure electoral system in india. the method used descriptive qualitative. the results indicate that a virtual voting system is environmentally friendly and is considered a resource-saving way for the election. it is because minimizes errors and increases voter participation through convenient virtual voting. in conclusion, a virtual voting system can develop an aadhar based advanced electronic voting machine (evm), which helps in a free and fair way of conducting elections. article history: ___________________ keywords: virtual voting system, electronic voting machine (evm), covid-19 1. introduction election is an activity for selecting regional leaders for countries that adhere to a democratic system. the electoral system is generally carried out by filling out the ballots of one of the candidates by the community. if one of the candidates gets the highest number of ballots, that candidate will be inducted to become the next leader, both at the state and regional levels within a country (loke et al, 2020). meanwhile, according to (roopak & sumathi, 2020) currently election activities have been developed using technology assistance. this is because the concept of an efficient election using virtual has begun to be applied, to assist international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(1) (2021) 77-82 received 18 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 kumar et al. virtual voting system| 78 the community in giving voting rights without being effective and in accordance with the location of the community. this is in line with (kathan et al, 2020) who said that the virtual election system is a tool used to provide certainty about the voting system by using biometric nuances and citizen vid (virtual id) obtained from the aadhar database to complete the voting and moreover using serious signature as key for voice encryption. the aadhaar card voting system is a voting system with citizen fingerprints. the qr code aadhaar will provide detailed information about the biodata of citizens who have voted (srinivasu & rao, 2018; thamizharasan & geetha, 2017). election plays important role in the democracy of india by giving chance to people to appoint their representatives from their respective constituencies. the virtual voting system using biometric can bring revolution by digitizing the process and by intensifying the trust and confidentiality of people who are keen to save their time, energy, and privacy and are friends with the virtual world of the internet. the need for a virtual voting system is to enable people to participate in the election from their homes in the pandemic era. in addition, the ineffective and inefficient traditional system can be minimized by the virtual voting system. other benefits of virtual voting system are (1) provide uniqueness and increases security through biometric, (2) easy to monitor and will require less manpower, (3) promote environmentally friendly election through the paperless system, (4) each vote in virtual voting system with a fingerprint is verifiable. the previous study has explained about virtual voting system (freitas & marie macadar, 2017) stated that e-voting system in brazil has many beneficial compares to the traditional system. unfortunately, there are some concerns about auditability, security, and costs of e-voting systems. (adamescu, 2021) stated that online voting in estonia could increase turnout and facilitate the category of voters who are less fortunate (older, foreign, and disabled) to express their political preferences. this support by (goodman & stokes, 2020) describes that online voting, in particular, has been found to increase voter turnout especially for marginalized populations, and provide counting efficiency in canada. evoting in indonesia can reduce errors and speed up the counting process are some of the benefits of electronic voting according to (sensuse, 2020). in addition, the e-voting system needs to provide proper arrangements in terms of discretionary procedures and this system adds the number of voters in pakistan (baidani, 2018). in addition, the previous research was conducted in other countries. therefore, the research about virtual voting systems needs to be performed based on specified areas such as india. it is because every country has a different election system. in conclusion, a virtual voting system can develop an aadhar based advanced electronic voting machine (evm) which helps in a free and fair way of conducting elections. additionally, the citizen can be sure that they can choose their leaders and realize their rights in a democracy 2. literature review we made the prototype by using the tools below: 1) java the program is developed on one machine and it is possible off on another 79 | international journal of informatics information system and computer engineering 2(1) (2021) 77-82 machine and get more secure java exploits. 2) mysql dbms it permits combination, extraction, manipulation, and organization of knowledge within the voter’s info. its platform is freelance and thus may be enforced and used across many like windows, linux servers and is compatible with varied hardware mainframes. it is quick in performance, stable, and provides business price at a coffee price. 3) xampp server xampp helps an area host or server to check its website and purchasers via computers and laptops before emotional it to the most server. it is a platform that furnishes an appropriate setting to check and verify the operating of comes supported apache, perl, mysql info, and php through the system of the host itself. 4) netbeans ide 8.0 the netbeans ide is an associate triumph integrated development setting offered for windows, mac, linux, and solaris. the netbeans project consists of associate ascii text file ide associated an application platform that modifies developers to speedily produce internet, enterprise, desktop, and mobile applications exploitation the java platform, moreover as php, javascript and ajax, groovy and grails, and c/c++. the technology used in the devices are shown in table 1. 3. results and discussion the new legal system can be enforced, victimization logins that require scanning of the candidate's fingerprint and name. this internet application supports all browsers. eligible voters can have their name, fingerprint, and alternative details in government information in each state or district as a visible match. thus, the fingerprint scanner can ensure that only legitimate voters will fake their votes. this app also ensures that the option is anonymous, when they log in each user is assigned a single and random id, which cannot have any ties to any user details, so there is no point in specifying that the user voted for that candidate. the focus is more on knowledge visual illustrations, and no free links are used, the interface is made as easy as possible with only basic functionality. the steps to use the device are: 1) scan your fingerprint and the application can match it with the info at the server. if the match is booming, the user is mechanically switched to succeeding voting window. 2) the ballot screen has all the logos and names of candidates standing for the post, the user simply must press the vote link next to his/her favorite candidate. if any user does not want to select any candidate for any reason, then he/she may directly log out exploitation the logout possibility. 3) auto-logout feature takes care of the remainder once a vote has been placed and the main login screen is restored. kumar et al. virtual voting system| 80 in this, we have made a system, which is fully web protocolled. the voter has been registered before voting so he/she can vote on voting day. they can vote wherever they are living and they do not need to come to hometown for voting. they have been given a password. hence, when voting they have to register their aadhar number and then it will scan their fingerprint, which will be encrypted in the form of a hash code. when the captcha and fingerprint are verified then the process will continue. if the process fails then the user cannot vote and if the data matches, then the process will continue. the project is designed to implement a virtual voting system using biometrics which eliminates the drawbacks of traditional voting systems in india. the benefits of virtual voting system: • online voting is an environmentally friendly and resourcesaving way to participatepaper voting is expensive and polluting. one of the advantages of online voting is its low resource requirements: compared with mail voting, online voting reduces co2 emissions by 98%. by switching to the internet, make environmentally friendly choices and save paper, printing, and transportation costs. you can also save staff and time by choosing online voting, and skip the tedious manual counting process. • minimize errors by casting off the usage of bodily put up and guide vote counting you may keep away from resultdistorting errors inclusive of lack of balloting files and miscounted votes. automatic vote counting with polyas online voting lets you get the right of entry to outcomes rapidly after the election. results also can be established through the usage of an outside tool. increase voter participation through convenient online voting: • voting should be hassle-free and easily accessible. reduce electoral barriers and provide safe online voting for eligible voters to increase voter turnout. the device can be used to: 1. create a secure online voting platform where authenticity of votes, voters and candidates is ensured using validated code on the backend. 2. improve voter identification as biometric features cannot be shared 3. relieve the problem of queuing at several points during the voting period in the election. there are some characteristics of virtual voting as follows: 1. eligibility: only eligible voters are allowed to cast their ballots. 2. privacy: there is no affiliation between voter identification and tagged ballots 3. uniqueness: voters can only vote once. 4. completeness: no one can enter valid ballots and voter ballots cannot be changed, valid ballots are counted correctly. 5. justice: no one can change the voting results. 6. verifiability: voters can verify their vote. 81 | international journal of informatics information system and computer engineering 2(1) (2021) 77-82 7. enforceability: no voter can show what he or she voted for to others to prevent bribery. 8. efficiency: calculations can be performed in real time. 9. mobility: the voter can vote anytime and anywhere through the internet table 1. the technology used in the device of virtual voting system no technology used application 1 operating system windows 7 or above or linux 2 programming language java/j2ee 3 user interface html, css 4 client-side scripting javascript 5 server deployment glassfish server 4.6 6 database mysql 7 software jdk 1.7 8 web applications jdbc, jsp, servlets 4. conclusion the "virtual voting system" project was completed to develop an advanced aadhar based electronic voting engine (evm) that assists in the free and fair conduct of elections. the advent of this project will enable the hosting of a fair election in india. it will preclude illegal practices like rigging. the citizen can be sure that they can choose their leaders, thus exercising their rights in the democracy. this project not only covers all drawbacks of the traditional voting system but also provides additional security. since fingerprint is unique, it reduces the chance of invalid and multiple votes. the manufacturing of this project is simple and low-cost. references adamescu, v. c. elections during pandemics. is i-voting a viable solution?. baidani, m. m., ahmad, m., & ali, y. (2018). an extended framework of online electronic voting system for pakistan. international journal of computer science and emerging technologies, 2(2), 17-21. de freitas¹, j. l., & macadar, m. a. (2017). the brazilian electronic voting system: evolution and challenges. second in, 59. kumar et al. virtual voting system| 82 goodman, n., & stokes, l. c. (2020). reducing the cost of voting: an evaluation of internet voting’s effect on turnout. british journal of political science, 50(3), 11551167. khatal, m. s. b., musmade, m. v. r., waman, m. t. a., shinde, m. s. b., & vikhe, m. n. b. (2020). aadhaar base voting system using blockchain technology, 6(5), pp.161-163 loke, y. c., batcha, n. k., & ab ziz, n. s. b. n. (2020). blockchain-enabled election voting system. journal of applied technology and innovation (e-issn: 26007304), 4(4), 51. roopak, t. m., & sumathi, r. (2020, march). electronic voting based on virtual id of aadhar using blockchain technology. in 2020 2nd international conference on innovative mechanisms for industry applications (icimia) (pp. 71-75). ieee. sensuse, d. i., & pratama, p. b. (2020, august). conceptual model of e-voting in indonesia. in 2020 international conference on information management and technology (icimtech) (pp. 387-392). ieee. srinivasu, l. n., & rao, k. s. (2018). aadhaar card voting system. in proceedings of international conference on computational intelligence and data engineering (pp. 159-172). springer, singapore. thamizharasan, n., & geetha, a. (2017). integration of biometric sensor with aadhar for voting process. journal of environmental nanotechnology, 6(1), 19-22. 1 | international journal of informatics information system and computer engineering 2(1) (2021) 47-54 utilization of communication technology for business lilis puspitawati *, a nurhasanah**, a s khaerunnisa*** accounting departement, economic and business faculty, universitas komputer indonesia, indonesia e-mail: * lilis.puspitawati@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o advances in technology, computers, and telecommunications support the development of internet technology. currently, the it function in society has begun to change. initially, the it function as a means of exchanging electronic information to an application that is used to implement corporate business strategies such as marketing, sales, and customer service activities. the purpose of this study is to get empirical evidence of the benefits of using communication technology for business people. this research used a descriptive method by conducting a survey in the form of a questionnaire and conducting several further interviews. the results showed that using communication technology such as the digital platform can make it easier for sellers to market products and make it easier to interact with buyers. this research discusses that using communication technology such as the digital platform is one of the best business strategies. the conclusion obtained in this study is that the digital platform is the best solution to expand business networks and make it easier for sellers and buyers to make transactions. the benefit of doing this research is to find out the meaning and purpose of developing business communication by involving digital technology in marketing products or services. article history: ___________________ keywords: information system, communication technology, business, international journal of informatics information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 2(1) (2021) 47-54 received 13 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 puspitawati et al. utilization of communication technology for business| 48 1. introduction currently, advances in technology, computers, and telecommunications are growing and changing rapidly. almost every second, there is always a new invention to improve or perfect previous technology results. with the maturity of technology, various technology types have emerged, one of which is communication technology where all users can connect. however, business people change communication technology's function from a tool for electronic information exchange to a tool for business strategy applications such as marketing, sales, and customer service (soegoto, e. s., & huda, m. n. 2019). in addition, communication technology is currently being used for individual needs. for business actors, communication technology can be used to achieve a competitive advantage. meanwhile, for individual needs, communication technology is used for personal interests such as looking for something needed, such as looking for products, job vacancies, etc. nowadays, businesses without utilizing technology will not be able to progress and are threatened with bankruptcy. many business people use technology to support business progress and get the desired benefits. utilizing technology, one of which is communication technology in the business sector, has a big positive impact on the business nets that we build (soegoto, e. s., & wardhani, a. n. k. 2018). previous research by muller explained that the digital platform is one of the results of communication technology. the research shows that the digital platform's main potential is to reduce transaction costs, combine company strengths, and realize economies of scale and coverage. digital platforms present challenges such as lack of trust, competitive thinking, high coordination efforts, and classified information loss (müller, j m. 2019). this study aims to get empirical evidence of the benefits of use communication technology for business. the data in the study were collected by conducting interviews with business people in the city of bandung. this research refers to previous research conducted by (hagiu, a., & wright, j. 2015; arnold, et al., 2016), (gawer, a., & cusumano ma. 2014; müller, et al., 2018; koskinen, et al., 2019; soegoto, e. s., & akbar, r. 2018), and (ruggieri, et al., 2018), succeeded in providing empirical evidence that the use of digital platforms has a role in increasing business prospects for various types of businesses. 2. method this study used a descriptive method with a quantitative approach, in which a survey was conducted through a questionnaire with 17 respondents and conducted further interviews with several respondents. all respondents are universitas komputer indonesia students who have a business, which utilizes communication technology, namely the digital platform, to apply business strategies. the period for filling out the questionnaire and interviewing was conducted for 2 days, from 19 to 20 january 2021. 3. results and discussion 3.1. definition of digital platform digital platform is one result of the development of communication technology today. the digital platform is 49 | international journal of informatics information system and computer engineering 2(1) (2021) 47-54 a technology that enables companies to uniform, edit and distribute data on an unprecedented scale (yoo, henfridsson, & lyytinen, 2010). companies use the digital platform as a prayer of one way to build a competitive advantage (parker, van alstyne, & choudary, 2016). by collecting, managing, and analyzing data, a unified platform, for example, partners, customers, and suppliers on one platform that serves the interests of multiple users (kiel, et al., 2017; hagiu, a., & wright, j. 2015; arnold, et al., 2016). it can open up new perspectives and new forms of interaction and relationships. they provide the basis to create new business models (xie, et al., 2016). in the study, digital platforms are understood as "products, services, and technology that are arranged in a common structure. where companies can create derivative products, services, and technology" (gawer, a., & cusumano ma. 2014); in this respect, digital platforms are different from traditional technology platforms. such technology is usually characterized by providing several products and services by the platform provider to its customers (müller, j m. 2019). the combination of several customer groups can also serve partly as a provider; for example, data for other customers and their interconnection in real-time is not visible on the platform's traditional technology (hagiu, a., & wright, j. 2015; gawer, a., & cusumano ma. 2014; müller, et al., 2018). there are three types of digital platforms as follow (koskinen et al., 2019): 3.1.1 transaction platform type the transaction platform is a digital platform concentrated on transactions or commonly called a multi-sided market (exchange platform). this type of platform's main purpose is to facilitate transactions between different organizations, entities, and individuals, such as connecting sellers and buyers, drivers and passengers, and others. this type of platform will be very useful in reducing transaction costs, where this platform can allow groups of users to search until they can find each other easily, and overall can reduce some problems in the transaction process (soegoto, e. s., & akbar, r. 2018). 3.1.2 innovation platform innovation platforms provide a technology platform, often covering a common set of standards, by which an ecosystem of third parties can develop complementary products and services for resale to consumers and other businesses (soegoto, e. s., & akbar, r. 2018). the innovation platform provides third-party developer tools and resources that developers combine and use legacy methods to enable new applications for commercial or other types of use. 3.1.3 platform integration the integration platform combines aspects of two types of platforms: transaction and innovation platforms (soegoto, e. s., & akbar, r. 2018). the transaction and innovation platform's key points also apply to the integration platform and are therefore not discussed again. 3.2. how the digital platform works as previously explained, the digital platform collects, manages, and analyzes data. it is shown in fig. 1. it is shown that the digital platform collects data from customers, producers, and providers. it can be in the form of product data, services, and personal information. after all the data is obtained, the digital platform will manage and analyze the data until the puspitawati et al. utilization of communication technology for business| 50 digital platform can unite and connect all users by simply searching to easily connect, such as connecting sellers with buyers and others. 3.3 uses of the digital platform based on the questionnaires that have been distributed and in-depth interviews, there are uses for the digital platform that respondents felt by respondents, namely business actors, which is shown in table 1. fig. 1. how the digital platform works (www.bearingpoint.com) table 1. the use of the digital platform, according to the respondents usability description example facilitate communication between sellers and buyers by using digital platforms, we can communicate more quickly with our customers doing direct messages like on instagram, twitter, and others product marketing reach is wider can make products and services that the respondents have can be spread more widely respondent a is domiciled in a city in indonesia; using the digital platform. he can market his products throughout indonesia and even globally efficiency digital platform can automatically make efficient, both in terms of marketing and manufacturing by utilizing the features provided, business people can optimize their business strategy increase innovation with the increasing number of digital platforms users, the more types of products and services needed by users will lead to new business models masks, at this time, there are various types of masks that customers can choose according to their respective interests and needs http://www.bearingpoint.com/ 51 | international journal of informatics information system and computer engineering 2(1) (2021) 47-54 based on table 1, we can see the various uses that business people feel by using the digital platform as a medium for applying business strategies. 3.4. reasons for using the digital platform using the digital platform is currently an obligation for businesses and customers. according to the interviews conducted for business actors, the reason they use the digital platform is that most of them do not have their shops / offline stores. meanwhile, if they use the digital platform, they do not have to own offline stores to market products or services. besides, the customers prefer the digital platform because they can easily find products or something they need. 3.5. digital platform used and the features provided based on the results of questionnaires, it shows several digital platforms that business people use for their business interests. it is shown in fig. 2 based on fig. 2, it is known that 70.6 % of users used digital platforms in instagram. in addition, there are 0% on twitter, 5.9% on youtube, 47.1% on whatsapp, and 23.5% on facebook. from these digital platforms, we compare the features available on the top three platforms selected as a consideration for which platforms provide features that make it easier for business actors. it is described in table 2 as follows: then, we explained the results of the level of respondent satisfaction with the digital platform. it is shown in fig. 3 as follows. fig. 3 shows that not all respondents feel that the digital platform is very useful, but no respondent feels that the digital platform is useless. therefore, further interviews were conducted with two respondents who felt less satisfied with the digital platform. furthermore, the result is that they feel that to run the digital platform at this time, besides being required the best business strategy, business actors must also face the existing challenges, one of which is building customer trust in the products/services they offer. fig. 2. digital platforms used by business actors fig. 3. the satisfaction level of business actors with the digital platform puspitawati et al. utilization of communication technology for business| 52 table 2. features of instagram, whatsapp, and facebook platform features usability instagram (account business) insight auto reply inbox filter stories live hashtag analyze who is logged into our business account so they can find out customer preferences accelerate the response to customers who ask for various information about our business mark several incoming messages to determine which ones should be reviewed and can filter out which messages have not been read so that they can maintain "engagement" between sellers and customers promote our business as attractive as possible, also build conversations with customers by using the question feature doing questions and answers or sharing other things directly where customers can see and respond immediately. useful for prospecting and looking for product inspiration, it can also show the character of our business whatsapp (business) profile automatic message message statistics label the profile contains our business information such as an address, email, business catalog, business description, and others set up automatic reply messages, so customers don't wait long review the number of messages sent, received, and read statistics group chats based on certain criteria (new customers, new orders, waiting for payment, orders completed) facebook inbox group appointment event job vacancy shopping insight there are several types of inboxes: messenger (interacting individually with a large scale reach), instagram direct (communicating with customers who are interested in our business on instagram), comments (including comments from facebook as well as instagram) provide a space to communicate with a group of people who are interested in our business allows customers to make appointments directly on facebook help promote our business to interested customers help recruit employees help share our services or products with customers and simplify the purchasing process shows statistics on interactions and actions taken by customers on our business accounts 53 | international journal of informatics information system and computer engineering 2(1) (2021) 47-54 4. conclusion based on research results, the digital platform most widely used by people every day is instagram, facebook, and whatsapp. therefore, the conclusion obtained in this research is that communication technology such as the digital platform is the best solution to expand business networks and make it easier for sellers and buyers to make transactions. references arnold, c., kiel, d., & voigt, k. i. (2016). how the industrial internet of things changes business models in different manufacturing industries. international journal of innovation management, 20(08), 1640015. gawer, a., & cusumano ma. (2014). industry platforms and ecosystem innovation. journal product innovative management. 31, 417–433. hagiu, a., & wright, j. (2015). multi-sided platforms. international journal of industrial organization, 43, 162-174. kiel, d., müller, j. m., arnold, c., & voigt, k. i. (2017). sustainable industrial value creation: benefits and challenges of industry 4.0. international journal of innovation management, 21(08), 1740015. koskinen k., bonina, c., & eaton b. (2019). digital platforms in the global south: foundations and research agenda. in international conference on social implications of computers in developing countries. 319-330. springer, cham. müller, j m. (2019). antecedents to digital platform usage in industry 4.0 by established manufacturers. sustainability, 11(4), 1121. müller, j.m., pommeranz, b., weisser, j., & voigt, k.i. (2018). digital, social media, and mobile marketing in industrial buying: still in need of customer segmentation? empirical evidence from poland and germany. industrial marketing management. 73, 70–83. ruggieri, r., savastano, m., scalingi, a., bala, d., & d’ascenzo, f. (2018). the impact of digital platforms on business models: an empirical investigation on innovative start-ups. management & marketing. challenges for the knowledge society, 13(4), 1210-1225. soegoto, e. s., & akbar, r. (2018). effect of the internet in improving business transactions with online market methods. in iop conference series: materials science and engineering, 407(1), 012051. soegoto, e. s., & huda, m. n. (2019). utilization of information technology as online business marketing media. in iop conference series: materials science and engineering, 662(3), 032018. puspitawati et al. utilization of communication technology for business| 54 soegoto, e. s., & wardhani, a. n. k. (2018). the role of information technology in online sales (online shopping). in iop conference series: materials science and engineering, 407(1), 012055. xie k., wu y., xiao j., & hu, q. (2016). value co-creation between firms and customers: the role of big data-based cooperative assets. information management. 53. 1034–1048. 1 | international journal of informatics information system and computer engineering 2(1) (2021) 55-64 diagnosis of covid-19 using chest x-ray sumit malik*, shivendra singh**, narendra mohan singh***, naman panwar**** noida institute of engineering and technology, india e-mail: *yaduyash@niet.co.in a b s t r a c t s a r t i c l e i n f o covid-19 is also a wide spreading infective agent disease that infects humans. a clinical study of covid19 infected patients has shown that these kinds of patients are square measure principally infected from a respiratory organ infection when come in contact with this disease. chest xray (i.e., radiography) a less complicated imaging technique for identification respiratory organ connected issues. deep learning is that the foremost undefeated technique of machine learning, that provides helpful analysis to review an oversize quantity of chest x-ray pictures which may critically impact on screening of covid-19. throughout this work, we have taken the pa read of chest x-ray scans for covid-19 affected patients conjointly as healthy patients. we have used deep learning-based cnn models and compared their performance. we have equate resnext models and inspect their precision to investigate the model presentation, 6432 chest x-ray scans samples square measure collected from the kaggle repository. this work solely core on potential ways of cluster covid-19 infected patients. article history: ___________________ keywords: deep learning, radiography images, covid-19 classification, convolutional neural network (cnn) 1. introduction covid-19 is a severe disease and a large number of individuals lose their lives a day. this disease affects not solely one country, but the entire world suffered as a result of this virus disease. within the contemporary world, the entire world is full of covid-19 disease, and also the most significant factor is not any single country scientists will prepare a vaccine for constant. in march 2020, x-ray image international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(1) (2021) 55-64 received 16 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 malik et al. diagnosis of covid-19 using chest x-ray | 56 of healthy folks and covid-19 infected peoples were obtainable on-line in numerous repositories like git hub, kaggle for analysis (wang y, et al. 2020). covid-19 is an epidemic disease that threatens humans at a world level and changed into an outbreak. the novel coronavirus disease came 1st as a strep throat, and suddenly folks faced issue in respiration. this infection is following a series method (roosa k, et al. 2020) that transfers from one person to a different when returning in grips with covid-19 infected persons. hospital workers, nurses, doctors, and clinical facilities play a vital role within the designation of this epidemic. a lot of technique area unit applied to cut back the result of coronavirus. medical imaging is additionally a way of analyzing and predicting the consequences of covid-19 on the anatomy. in this, healthy folks and covid-19 infected patients are often analyzed in parallel with the assistance of chest x-ray image. for contributing to an analysis of covid-19, we have a tendency to collected uploaded knowledge of xray image of healthy and covid-19 infected patients from totally different sources and applied 3 different models (inceptionv3, xception, and resnext) (kanne jp, 2020). the analysis of this collected knowledge is finished with the assistance of cnn, a machine learning tool. this work mainly focuses on the utilization of cnn models for classifying chest x-ray images for coronavirus infected patients. we have tried to draw a parallel to the previous add the sector and appearance for potential models of the task, which may be assessed additional to prove their quality in sensible eventualities. deep learning based mostly models (and a lot of specifically convolutional neural networks (cnn)) are shown to exceed the classical ai approaches in most of system vision and and medical image analysis tasks in recent years. this paper, additional classified within the completely different sections. mentioned numerous researchers’ views in analyzing the impact of the covid-19 disease on countries and humans. dataset used and model formulation, completely different matrices and algorithms used. further, the analysis of leads to terms of training and testing with confusion matrices for models used next. 2. literature review following are a few literature reviews papers we discovered and read. these papers contain similar implementations and ideas whereas they have some or the other limitations. we received the following papers in order to come to a few conclusions as well as define their limitations when compared to our approach: • afshar p, heidarian s, naderkhani f, oikonomou a, plataniotis kn, mohammadi, the authors projected a framework model supported capsule networks to diagnose covid-19 (i.e., covid-caaps) illness with the assistance of x-ray pictures. during this projected work, many convolution layers and capsules square measure wont to overcome the matter of class-imbalance. in experimental analysis, as a result, they ended that the projected model shows accuracy 95.7%, whereas sensitivity is shown as 90% and specificity as 95.80% whereas put in a smaller sort of trainable parameters (afshar p, et al. 2020). • alqudah am, qazan s, alqudah, the author planned a hybrid system supported computing, that specially used 57 | international journal of informatics information system and computer engineering 2(1) (2021) 55-64 machine learning and deep learning algorithms (i.e., convolutional neural network (cnn) apply softmax classifier). the planned system is specially enforced for police investigation covid-19 cases apply chest x-ray pictures (alqudah am, et al. 2020). • hassanien ae, mahdy ln, ezzat ka, elmousalami hh, ella, the authors counselled a deep primarily based methodology (with vector device classifier) for the detection of patients infected from covid-19 by apply x-ray pictures. this technique is useful to hospital doctors for early sleuthing the cases of covid-19 infected patients. they realize 97.48% accuracy of the planned model for respiratory organ classification with the assistance of various matrices parameters (hassanien ae, et al. 2020). • ilyas m, rehman h, nait-ali, the authors mentioned the various methodologies used for covid-19 illness detection and challenges sweet-faced. they additionally aforementioned that associate in nursing automatic methodology for detective work the covid-19 virus ought to be developing to forestall the spreading of the disease through contact. then, they analysed completely different chest x-rays for the detection of respiratory disorder and terminated that it's laborious to predict that covid-19 causes respiratory disorder or the other symptom's area unit liable for this (ilyas m, et al. 2020). • ozturk t, talo m, yildirim ea, baloglu ub, yildirim o, acharya ur, the authors projected a model that mechanically detects the covid-19 with the assistance of chest x-ray pictures. the planned model is employed to convey correct medicine on 2 completely different classification models (i.e., binary and multi-class). they apply the darknet model to classify the period of time object detection method (ozkaya u, et al. 2020). 3. proposed system overview the given model shows that how the proposed system will work, all the steps we take to implement it and use it to predict the data based on the performance of it (see fig. 1). fig. 1. steps implementation system 4. feature extraction 4.1. methodology 1. dataset the source where we have collected the dataset was the kaggle repository, which contains chest x-ray scans of covid-19 affected, normal. the collected dataset have 6432 total chest x-ray images. this data set is further divided into training (i.e., 5467) and validation (i.e., 965) set of normal, covid, and pneumonia. in the training set, 1345 is normal, 490 are covid, and 3632 is pneumonia. in the validation phase, 238 samples of a normal case, 86 covid, and 641 of pneumonia were considered for this analysis, were considered for this analysis. the scans were scaled down (table 1). malik et al. diagnosis of covid-19 using chest x-ray | 58 table 1. displays the data distribution for training and testing the data 128 × 128 to assist the quick coaching of our model. 2. model formulation the dataset to stop over-fitting. the augmentations enclosed rotation, zoom, and sharing of pictures. the information was then shuffled to generalize the model and cut back over-fitting. after this, the ready dataset was wont to train the projected model (wang s, et al. 2020). for higher analysis, 3 completely different models are enforced, and so their performance was compared to calculate the accuracy. within the given models, we have a tendency to enforced leakyrelu activation rather than the originally used relu activation operate, which makes it as a unique technique. this method helps to hurry up the training and conjointly avoids the matter of dead neurons (i.e., the relu neurons become inactive because of zero slopes). figure one shows the projected model for chest x-ray image analysis (see fig. 2). fig. 2. the architecture of cnn model 3. inception net v3 inception internet v3 may be a cnn based mostly network for classification. it's forty-eight layers deep and uses beginning modules, that includes of a concatenated layer with 1 × 1 3 × 3 and 5 × 5 convolutions. by doing this, we will decrease the number of parameters and increase the coaching speed. it's additionally noted as google net design. 4. xception net it is a modification of the origination internet. during this model, the origination modules square measure replaced with depth wise divisible convolutions. its parameter size is analogous to the origination internet (chollet f, 2017). however, it performs slightly higher as compared to the origination internet. 5. resnext resnext is an associate degree extension design of the deep residual network. during this model, the quality remaining blocks square measure replaced with one that leverages a split remodel merge strategy employed in the inception models. 4.2. algorithm step 1:preprocess image i.e. image = x preprocess apply (we have utilized keras data generator for this goal) 1. reshape image (x) to (128, 128, 3) 2. random rotation range is 10° 3. horizontal flip is true 4. zoom range is 0.4 step 2: applying the picture to firstly input of the pre-trained model step 3: get 59 | international journal of informatics information system and computer engineering 2(1) (2021) 55-64 the output of the last convolution layer of the given model. step 4: flatten size with reducing n size to n-1. step 5: put a dense layer units is equal 256 for inception net and xception net units is equal 128 for resnext z=w (a+bz)=w (a+b) step 6: apply a=leakyrelu(z)a=leakyrelu(z) step 7: apply dense layer for abstract thought z=w (a+bz)=w (a+b) step 8: apply softmax for classification 5. experimental result 5.1. matrices used for result evaluation thi proposed model has been evaluated with the help of different parameters such as precision, recall, f1 score, its accuracy, sensitivity, and specificity as shown in equation below. precision = xception net: it is a modification of the origin internet. during this model, the origin modules area unit replaced with depth wise dissociable convolutions. its parameter size is comparable to the origin internet, however it performs slightly higher as compared to the origin internet (tables 2 and 3). table 2. f1-score for training dataset for xception net model table 3. f1-score for testing dataset for xception net model (a) confusion matrix of train data of xception model. (b) confusion matrix of test data of the xception model (see fig. 3) soft max( ∑ malik et al. diagnosis of covid-19 using chest x-ray | 60 fig. 3. training and testing data of the xception model shows the confusion matrix. inception net v3: it is a state of the art convlution neural network for classification. it's 48 layers deep and uses beginning modules, that contains a concatenated layer with 1 × 1 3 × 3 and 5 × 5 convulsions. doing this, we are able to decrease the quantity of parameters and increase the training speed. it's conjointly named as google net design (tables 4 and 5). table 4. depict the f1-score on training dataset for the inception v3 model. table 5. depict the f1-score on testing dataset for the inception v3 model. resnext: this design is to associate in nursing extension of the deep residual network. during this model, the quality of remaining blocks area unit replaced with one that leverages a split rework merge strategy utilized in the inception models (tables 6 and 7). (a): confusion matrix of train data of inception v3, (b) confusion matrix of test data of inception v3 (see fig. 4) 61 | international journal of informatics information system and computer engineering 2(1) (2021) 55-64 fig. 4. training and testing data of inception v3 model shows the confusion matrix. table 6. f1-scores for the training dataset for resnext model table 7. f1-scores for the testing dataset for resnext model (a). confusion matrix of train data of resnext model. (b): confusion matrix of test data of resnext model (see fig. 5). in the analysis of result, typical chest xray pictures are compared with covid19 affected people. inception net v3, xception net and res next are analysed supported accuracy matrices. the results were then contrasted to work out the simplest model. although the model predictions are very high, we suggest validating the execution using future updates on the dataset. because of the limited amount of information, the model is trained only on 1560 samples (see figs. 6 and 7). malik et al. diagnosis of covid-19 using chest x-ray | 62 fig. 5. training & testing data of the resnext model shows the confusion matrix. fig. 6. mockup upload image fig. 7. mockup results 6. advantages the following advantages of the implemented method or system are mentioned below: 1. the cost of detecting covid19 using chest x-ray is low as compare to rtpcr kit. 2. through rtpcr kit it takes 24 hours for the result but if we started the test through chest x-ray it takes less time. 7. conclusion a covid-19 pandemic is also growing very fast. with the ever-increasing range of cases, bulk testing of cases fleetly might even be needed. throughout this work, we have a tendency to experimented with multiple cnn models in an attempt to classify the covid-19 affected patients mistreatment their chest x-ray scans. further, we have a tendency to over that out of these 3 models, the xception web has the best performance and is suited to be used. we have with success classified covid-19 scans and it describes the likely range of applying like procedure within the close future to automate analysis tasks. the high accuracy obtained could also be an 63 | international journal of informatics information system and computer engineering 2(1) (2021) 55-64 explanation for concern since it's going to be a result of overfitting. this will be verified by testing it against new data that's made public shortly. within the future, the big dataset for chest x-rays is often considered to validate our proposed model thereon. it's also advised to consult medical professionals for any practical use case of this project. we develop a perfect detection mechanism however solely analysis concerning attainable economically possible ways in which to fight this unwellness. such strategies might even be pursued for additional analysis to prove their real case implementation. references afshar, p., heidarian, s., naderkhani, f., oikonomou, a., plataniotis, k. n., & mohammadi, a. (2020). covid-caps: a capsule network-based framework for identification of covid-19 cases from x-ray images. pattern recognition letters, 138, 638-643. alqudah, a. m., qazan, s., & alqudah, a. (2020). automated systems for detection of covid-19 using chest x-ray images and lightweight convolutional neural networks. chollet, f. (2017). xception: deep learning with depthwise separable convolutions. in proceedings of the ieee conference on computer vision and pattern recognition (pp. 1251-1258). mahdy, l. n., ezzat, k. a., elmousalami, h. h., ella, h. a., & hassanien, a. e. (2020). automatic x-ray covid-19 lung image classification system based on multi-level thresholding and support vector machine. medrxiv, 2020-03. ilyas, m., rehman, h., & naït-ali, a. (2020). detection of covid-19 from chest x-ray images using artificial intelligence: an early review. arxiv preprint arxiv:2004.05436. kanne, j. p. (2020). chest ct findings in 2019 novel coronavirus (2019-ncov) infections from wuhan, china: key points for the radiologist. özkaya, u., öztürk, ş., & barstugan, m. (2020). coronavirus (covid-19) classification using deep features fusion and ranking technique. in big data analytics and artificial intelligence against covid-19: innovation vision and approach (pp. 281295). springer, cham. roosa, k., lee, y., luo, r., kirpich, a., rothenberg, r., hyman, j. m., ... & chowell, g. b. (2020). real-time forecasts of the covid-19 epidemic in china from february 5th to february 24th, 2020. infectious disease modelling, 5, 256-263. wang s, kang b, ma j, zeng x, xiao m, guo j, cai m, yang j, li y, meng x, xu b (2020) a deep learning algorithm using ct images to screen for corona virus disease (covid-19) malik et al. diagnosis of covid-19 using chest x-ray | 64 wang, y., hu, m., li, q., zhang, x. p., zhai, g., & yao, n. (2020). abnormal respiratory patterns classifier may contribute to large-scale screening of people infected with covid-19 in an accurate and unobtrusive manner. arxiv preprint arxiv:2002.05534. 13 | international journal of informatics information system and computer engineering 2(1) (2021) 13-22 crowd detection using yolov3-tiny method and viola-jones algorithm at mall selvia lorena br ginting*, hanhan maulana**, riffa alfaridzi priatna***, deran deriyana fauzzan****, devidli setiawan***** *center for artificial intelligence technology, universitas kebangsaan malaysia, malaysia ** japan advanced institute of science and technology ***,****,*****departemen teknik informatika, universitas komputer indonesia, indonesia e-mail: *selvialorena@yahoo.com a b s t r a c t s a r t i c l e i n f o indonesia is one of the countries affected by covid-19 which is spreading quite fast. lately, the surge in covid19 cases in indonesia is quite high, due to the lack of public awareness of the current health protocols, such as avoiding crowds and keeping a distance. the purpose of this study is to reduce crowds that occur in places with a high risk of crowding, for example in mall. detection is done by using closed circuit television (cctv) in the mall and using the yolov3-tiny method and the violajones algorithm to detect the crowd. to support the research, we use the method of literature study and field observation at cimahi mall as one of the malls in the area of bandung raya. the results show that to reduce the number of crowds that occur in the mall, crowd detection must be carried out using the yolov3-tiny method and the viola-jones algorithm, and a warning system is given if a crowd is detected in the place. the main concept of this system is crowd detection and warning if there is a crowd located on cctv in the mall. in our opinion, when this system is running in malls that occur in indonesia, the number of positive cases of covid-19 in indonesia will decrease because there are no crowds. it can be concluded that this system exists as a precaution against the crowds that often occur today at the mall. prevention is done by detecting crowds and giving warnings if there is a crowd so that positive cases of covid-19 in indonesia will be reduced. article history: ___________________ keywords: crowd, detection, yolov3-tiny, viola-jones. international journal of informatics, information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 2(2) (2021) 13-22 received 10 nov 2021 revised 20 nov 2021 accepted 25 nov 2021 available online 26 dec 2021 selvia et al. crowd detection using yolov3-tiny method...| 14 1. introduction covid-19, also known as the deadly coronavirus, has already taken half of china's life. sars-cov-2 or 2019-ncov was first detected in the city of wuhan, china in december 2019 since the disease has spread exponentially. while covid19 patients on 28 february 2020 were 86,604 and 858,361 on 31 march 2020, the number increased to 2,086,477 on 15 april 2020 (huang et al., 2020). the government itself has notified indonesian citizens to self-isolate.covid-19, also known as the deadly coronavirus, has already taken half of china's life. sars-cov-2 or 2019ncov was first detected in the city of wuhan, china in december 2019 since the disease has spread exponentially. while covid-19 patients on 28 february 2020 were 86,604 and 858,361 on 31 march 2020, the number increased to 2,086,477 on 15 april 2020 (huang et al., 2020). in indonesia itself, this virus was first entered in march 2020 and the steps taken by the government were to enforce the policy of pemberlakuan pembatasan kegiatan masyarakat (ppkm) (nasution et al., 2020). then, in addition to ppkm, the government also provides education and several measures to prevent the spread of covid-19 to the public, including maintaining hand hygiene by washing hands, respiratory etiquette, and using masks, then implementing social distancing such as closing schools and universities, banning schools and universities. organizing large events and mass gatherings, restricting travel and public transportation, making the public aware of the dangers, being required to stay at home, and implementing selfisolation (aquino et al., 2020). however, there are still many indonesians who don't wear masks, don't keep their distance, hold big events. that way, more and more people congregate in one place. therefore, we need a technology to detect the distance in a crowd on closed circuit television (cctv). previous research said that social distancing is one of the steps to reduce the spread of covid-19 (courtemanche et al., 2020). to prevent crowds, object detection in the form of humans is carried out to detect whether they keep their distance or not by comparing the pixel values between two or more objects (parameswaran et al., 2021). detection is carried out using the yolov3 method to obtain stable fps and accurate results (punn et al., 2020). yolov3-tiny is used to provide lightening the performance of the processing device used so that it can be used in the long term (li et al., 2020). the viola-jones algorithm is strong enough to detect objects in real-time so that the results produced are also quite fast (jadhav & lahudkar, 2021). therefore, from this research, we can see that by preventing crowds, we can reduce the spread of covid-19, especially in indonesia. from this research, we can also combine the yolov3-tiny method and the viola-jones algorithm to produce fast, accurate data processing, stable fps, but not consuming large hardware resources. the purpose of this study is to reduce crowds that occur in places with a high risk of crowding, for example in malls. detection is done by using cctv in the mall and using the yolov3-tiny method and the viola-jones algorithm to detect the crowd. to support the research, we use the method of literature study and field observation at cimahi mall as one of the malls in the area of bandung raya. 15 | international journal of informatics information system and computer engineering 2(1) (2021) 13-22 2. method the method used in this research is the descriptive quantitative method, which is a method used to provide an overview of the actual situation (jalinus, n., & risfendra, 2020). the data collection technique that we use is passive observation, namely by observing every crowd that occurs within cctv coverage at cimahi mall. to perform the detection, we use the yolov3-tiny method and the viola-jones algorithm because these methods and algorithms can be used in the long term, giving the best and fast results (punn et al., 2020; li et al., 2020; jadhav & lahudkar, 2021). however, for the initial version and because the pandemic is still ongoing, we are prototyping the tool that we will build. the prototype is made so that it can reflect the expected results of this system if later it has been implemented in malls in indonesia (günther et al., 2021). 3. results and discussion 3.1. system design this crowd detection system will use several hardware devices such as cctv in the mall as input for this detection system. raspberry pi 4 as a tool to process input from cctv and detect objects in the form of humans. then, the system will decide whether there are two or more objects that are close to each other or not, if there are objects that are close to each other, the raspberry pi will issue an output through the buzzer and speaker which will be placed under the cctv to give a warning to the object that is close to each other. in addition to providing output through the buzzer and speaker, the raspberry pi will also be directly connected to the monitor in the cctv monitoring area and will provide a display of object detection. the circuit of the detection system is shown in fig. 1. for the output on the monitor that located at the cctv monitoring area, the system will indicate which a human object with a box is. the system will distinguish which objects do not keep their distance by giving two or more objects with a red box, whereas if the objects keep their distance well, the system will give the object a green box. the display on the monitor is shown in fig. 2. fig. 1. circuit of the detection system selvia et al. crowd detection using yolov3-tiny method...| 16 fig. 2. detection result in monitor 3.2 developing the initial prototype in the early part of the prototyping process, we performed detection using the webcam on the laptop and immediately ran the detection program on the laptop as well. for the prototype of this detection system, it will only provide a warning in the form of output in the terminal of the visual studio code application. the detection display on the laptop is the same as the one that will appear on the monitor in the cctv monitoring area. however, the difference lies in the output given, the following is the output given during this prototyping (fig. 3). fig. 3. detection result in visual studio code terminal 3.3 testing the prototype the test results are still carried out with a prototype. tests are carried out to find out how effective this detection system is to detect humans and to check whether there are two or more humans who are close to each other. if there is such a condition, the system will warn the two or more humans to immediately keep their distance, and if they do not comply with the order, they will be followed up by the covid-19 officer unit at the mall. after testing 40 times on prototype that was made only using the viola-jones algorithm, the detection results show that 93.24% can detect humans and 75% of them keep their distance from each other. the detection is said to be successful if it can detect an object in the form of a human being in the camera area, while the detection is said to fail if there is an object in the form of a human in the camera area, but the detection system cannot detect it. accuracy is obtained from the comparison of the number of successful detections with the number of successful and failed detections. the processing time results are obtained from how long the system can detect several objects in 1 image. the result of whether the object keeps its distance or not is obtained from the detection between several objects. if there are 2 or more objects that are detected close to each other, then the result is that a violation of social distancing is detected. the test results only with the viola-jones algorithm are shown in table 1. table 1. result test only using viola-jones algorithm [7] no image count human an image accuracy processing time (second) is maintain a distance? success failure 1 image 1 8 100% 10.2 yes 2 image 2 6 100% 2.7 yes 3 image 3 4 100% 4.2 yes 17 | international journal of informatics information system and computer engineering 2(1) (2021) 13-22 no image count human an image accuracy processing time (second) is maintain a distance? success failure 4 image 4 5 100% 3.1 yes 5 image 5 11 100% 4.2 yes 6 image 6 24 3 88.88% 12.2 no 7 image 7 30 1 93.70% 13.6 no 8 image 8 3 100% 3.6 yes 9 image 9 2 100% 2.4 yes 10 image 10 8 2 80% 14.5 yes 11 image 11 12 100% 3.7 yes 12 image 12 36 2 94% 5.9 no 13 image 13 5 100% 3.3 yes 14 image 14 4 100% 2.4 yes 15 image 15 12 100% 3.7 yes 16 image 16 8 100% 2.9 yes 17 image 17 5 100% 2.3 yes 18 image 18 24 3 88.88% 4.5 no 19 image 19 3 100% 2.3 yes 20 image 20 24 1 96% 4.9 no 21 image 21 4 100% 1.4 no 22 image 22 0% 1.2 yes 23 image 23 5 100% 1.5 yes 24 image 24 1 100% 2.4 yes 25 image 25 3 100% 7 yes 26 image 26 1 100% 1.3 yes 27 image 27 0% 1.2 yes selvia et al. crowd detection using yolov3-tiny method...| 18 no image count human an image accuracy processing time (second) is maintain a distance? success failure 28 image 28 5 100% 1.4 no 29 image 29 1 100% 22.2 yes 30 image 30 3 100% 3.3 yes 31 image 31 1 100% 1.8 yes 32 image 32 1 100% 1.4 yes 33 image 33 5 100% 2.2 yes 34 image 34 3 100% 1.1 yes 35 image 35 10 100% 11.5 yes 36 image 36 50 1 98% 31.5 no 37 image 37 2 100% 2.1 yes 38 image 38 31 1 96.80% 13.3 no 39 image 39 1 100% 1.7 yes 40 image 40 28 2 93.30% 4.9 no accuracy = 93.24% 75% after combining the yolov3-tiny method and the viola-jones algorithm, the results obtained from this test show that 96.32% can detect humans and 85% of them keep their distance from each other. the detection is said to be successful if it can detect an object in the form of a human being in the camera area, while the detection is said to fail if there is an object in the form of a human in the camera area, but the detection system cannot detect it. accuracy is obtained from the comparison of the number of successful detections with the number of successful and failed detections. the processing time results are obtained from how long the system can detect several objects in 1 image. the result of whether the object keeps its distance or not is obtained from the detection between several objects. if there are 2 or more objects that are detected close to each other, then the result is a violation of social distancing. the test results using the yolov3-tiny method and the viola-jones algorithm are shown in table 2. 19 | international journal of informatics information system and computer engineering 2(1) (2021) 13-22 table 2. result test using yolov3-tiny method and viola-jones algorithm no image count human an image accuracy processing time (second) is maintain a distance? success failure 1 image 1 8 100% 8.2 yes 2 image 2 6 100% 1.4 yes 3 image 3 4 100% 3.7 yes 4 image 4 5 100% 2.6 yes 5 image 5 11 100% 3.1 yes 6 image 6 26 1 96.30% 11.2 no 7 image 7 31 100 % 13.6 no 8 image 8 3 100% 2.4 yes 9 image 9 2 100% 2.1 yes 10 image 10 9 1 90% 12.5 yes 11 image 11 12 100% 3.3 yes 12 image 12 36 2 94% 5.9 no 13 image 13 5 100% 3.3 yes 14 image 14 4 100% 2.4 yes 15 image 15 12 100% 3.7 yes 16 image 16 8 100% 2.9 yes 17 image 17 4 1 80% 4.3 yes 18 image 18 26 1 96.30% 3.2 no 19 image 19 3 100% 2.3 yes 20 image 20 24 1 96% 4.2 no 21 image 21 4 100% 1.1 no 22 image 22 2 100% 1.5 yes 23 image 23 5 100% 1.8 yes selvia et al. crowd detection using yolov3-tiny method...| 20 no image count human an image accuracy processing time (second) is maintain a distance? success failure 24 image 24 1 100% 2.4 yes 25 image 25 3 100% 7 yes 26 image 26 1 100% 1.3 yes 27 image 27 0% 1.2 yes 28 image 28 5 100% 1.4 no 29 image 29 1 100% 22.2 yes 30 image 30 3 100% 3.3 yes 31 image 31 1 100% 1.8 yes 32 image 32 1 100% 1.4 yes 33 image 33 5 100% 2.2 yes 34 image 34 3 100% 1.1 yes 35 image 35 10 100% 11.5 yes 36 image 36 51 100% 31.5 no 37 image 37 2 100% 2.1 yes 38 image 38 32 100% 13.3 no 39 image 39 1 100% 1.7 yes 40 image 40 30 100% 4.9 no accuracy = 96.32% 85% the process of producing al2o3 nanoparticles using the precipitation method is carried out with several instruments using industrial scales that can be obtained commercially and economically. suppose the production is carried out 960 times a year. in that case, 6.9 tons of al2o3 nanoparticles will be produced, requiring 33.00 tons of aluminum chloride hexahydrate, 60.48 tons of ammonium hydroxide, 12.21 tons of tween-80, and 200 tons of ethanol. the total price required in a year for production is 203,364.31 usd with annual sales of 348,000.00 usd, resulting in a profit of 144,635.69 usd per year. 21 | international journal of informatics information system and computer engineering 2(1) (2021) 13-22 these advantages will be shown in an economic evaluation, and the value of the project will be shown over 20 years 4. conclusion this crowd detection system was created as a precautionary measure to prevent crowds and to reduce crowd levels that occur in places with a high risk of crowding, such as in malls. this system runs according to its purpose, which is to reduce the crowds that occur in malls so that the number of covid-19 cases in indonesia will continue to decrease so that later there will be no more covid-19 cases in indonesia. references aquino, e. m., silveira, i. h., pescarini, j. m., aquino, r., souza-filho, j. a. d., rocha, a. d. s., ... & lima, r. t. d. r. s. (2020). social distancing measures to control the covid-19 pandemic: potential impacts and challenges in brazil. ciencia & saude coletiva, 25, 2423-2446. courtemanche, c., garuccio, j., le, a., pinkston, j., & yelowitz, a. (2020). strong social distancing measures in the united states reduced the covid-19 growth rate: study evaluates the impact of social distancing measures on the growth rate of confirmed covid-19 cases across the united states. health affairs, 39(7), 12371246. günther, s., müller, f., hübner, f., mühlhäuser, m., & matviienko, a. (2021, june). actuboard: an open rapid prototyping platform to integrate hardware actuators in remote applications. in companion of the 2021 acm sigchi symposium on engineering interactive computing systems (pp. 70-76). huang, c., wang, y., li, x., ren, l., zhao, j., hu, y., ... & cao, b. (2020). clinical features of patients infected with 2019 novel coronavirus in wuhan, china. the lancet, 395(10223), 497-506. jadhav, r. r., & lahudkar, s. l. (2021). implementing a real time human detection and monitoring social distancing for covid-19 using vj algorithm and opencv. international journal, 10(2). jalinus, n., & risfendra, r. (2020). analisis kemampuan pedagogi guru smk yang sedang mengambil pendidikan profesi guru dengan metode deskriptif kuantitatif dan metode kualitatif. invotek: jurnal inovasi vokasional dan teknologi, 20(1), 37-44. li, t., ma, y., & endoh, t. (2020). a systematic study of tiny yolo3 inference: toward compact brainware processor with less memory and logic gate. ieee access, 8, 142931-142955. nasution, d. a. d., erlina, e., & muda, i. (2020). dampak pandemi covid-19 terhadap perekonomian indonesia. jurnal benefita, 5(2), 212-224. parameswaran, b., & sb, b. (2021). crowd detection for social distancing and safety violation alert based on image processing. selvia et al. crowd detection using yolov3-tiny method...| 22 punn, n. s., sonbhadra, s. k., agarwal, s., & rai, g. (2020). monitoring covid-19 social distancing with person detection and tracking via fine-tuned yolo v3 and deepsort techniques. arxiv preprint arxiv:2005.01385. 1 | international journal of informatics information system and computer engineering 1 (2020) 79 90 fog computing architecture for indoor disaster management asep id hadiana faculty of information and communication technology, universiti teknikal malaysia melaka, melaka, malaysia correspondence: e-mail: ahadiana@gmail.com a b s t r a c t s a r t i c l e i n fo most people spend their time indoors. indoors have a higher complexity than outdoors. moreover, today's building structures are increasingly sophisticated and complex, which can create problems when a disaster occurs in the room. fire is one of the disasters that often occurs in a building. for that, we need disaster management that can minimize the risk of casualties. disaster management with cloud computing has been extensively investigated in other studies. traditional ways of centralizing data in the cloud are almost scalable as they cannot cater to many latency-critical iot applications, and this results in too high network traffic when the number of objects and services increased. it will be especially problematic when in a disaster that requires a quick response. the fog infrastructure is the beginning of the answer to such problems. this research started with an analysis of literature and hot topics related to fog computing and indoor disasters, which later became the basis for creating a fog computingbased architecture for indoor disasters. in this research, fog computing is used as the backbone in disaster management architecture in buildings. mqtt is used as a messaging protocol with the advantages of simplicity and speed. this research proposes a disaster architecture for indoor disasters, mainly fire disasters. article history: received 14 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 online 09 sep 2018 ___________________ keywords: disaster management, fog computing, mqtt. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 1 (2020) 79 90 asep id hadiana. fog computing architecture for indoor disaster management| 80 1. introduction most people spent their time in daily life indoors, either consume their time in the office, campus, shopping centre, airport, and home. other studies also establish the average time of people spent their time indoors is around 90% for different contingents and time (klepeis et al., 2001). meanwhile, the advanced construction of the building nowadays is enormous, and it has a complex structure. the advance of building constructions possibly increases the complication for managing the urban disaster, which is fire. the concentrated human population in urban society has increased the potential for destruction, such as natural or human-made disasters. based on the statistics from 31 countries reported by the international association of fire and rescue services, in 2013, 2.5 fires are breaking out on average every minute and the total number of fires, deaths is more than 217. 000 worldwide. moreover, crowd behaviors of human beings, which are common occurrences in confined spaces, can have a significant impact on emergency evacuation. destructive crowd behaviors such as pushing and trampling caused by congestion may block evacuees and induce serious fatalities and injuries, sometimes even more severe than losses caused by fire (desmet et al., 2013). hence, people who are not acquainted with it may have difficulty to navigate through the building. this kind of emergency situation requires people to do rapid evacuation inside the building. in emergency scenarios the fast conduction of victims to exit and the precise monitoring of the rescuers position are both important to reducing deaths and injuries. location positioning systems can be used to automatically gather the positioning of rescuers and victims (bastos et al., 2015). the lack of real-time, coordinated responses, driven by integrated decision making facilities based on information collected by firstrespondents acting on a crisis site, is a critical challenge in emergency management (palmieri et al., 2016). disaster management is a sort of integration and coordination of actions required preserving, improving and preparing the capability to respond to disasters or natural disasters caused by humans. disaster management involves various tools and is also an initiative various area, of various parties. a cooperation and coordination structure is necessary to include multiple partners and players in disaster management. adoption of new techniques can reduce the chances of losing human lives as well as harm to large scale infrastructures due to both human-made and natural disasters. iot, which enables interconnection among devices with diverse performance, is a solution for disaster management. employing intelligence applications and data analytics, iot-enabled tragedy management methods used for warning about the mishap. since the impact of any tragedy is tremendous, the iot-enabled disaster management system can be implemented to obtain possible rescue operations and the victim (ray et al., 2017). localization expedite the healing procedure and helps in disaster control that is effective. among the issues of disasters is obtaining information what's their place from the location and whether they're secure or not. localization will help in these situations by providing 81 | international journal of informatics information system and computer engineering 1 (2020) 79 90 them with aid in situations like the consumer and offering the location of the people. likewise in the event of a flame or some other calamity within a environment, the user places can be obtained by the rescue group throughout the system that may be used for functioning at the place (zafari et al., 2019) (see fig. 1). fig. 1. emergency navigation systems (bi & gelenbe, 2019) since hazard sensors have gradually become the norm for public buildings, various wireless sensor network-based algorithms have proposed to provide dynamic evacuation plans in a real-time fashion. because of constraints in battery life capacity and processing capacity of a wireless sensor system, these algorithms are typically made in an efficient fashion. the rapid development of cloud computing technology in the last decade has made much research carried out by leveraging the enormous computing power of cloud servers to design more sophisticated and accurate algorithms to guide disaster evacuation. due to the high processing capacity, storage that is big degree, and interoperability, cloud computing systems has the capability and has come to be a tech. in comparison wsn-based counterparts emergency response methods are flexible because of the removal of the time to establish a wsn along with the maintenance, for example battery changes (gelenbe & bi, 2014). cloud computing is utilized to store and process the enormous amounts of information created by increasing iot devices. however, transmission delay difficulty occurs in the cloud computing system environment because the huge quantities of data is saved in the cloud servers that are remote. (khan et al., 2015) introduces a way for context aware route planning through coupling of a multipurpose cloud-based method structure using a complex indoorgml data model, and it's an impending ogc standard for representing and exchanging semantics, geometry, and topology advice of in-door 3d construction models. additional due to the immense dependence on connectivity along with regular power supply of cloud computing, the use of cloud services could be problematic throughout the occurrence of crises at which both energy and communication networks will crack down. efficient and easy usage of cloud infrastructure, particularly during the disaster, is still among the key challenges although cloud computing delivers a better solution for disaster simulation and modeling. (ujjwal et al., 2019) have shown that cloud-based solutions are somewhat more cost effective for running asep id hadiana. fog computing architecture for indoor disaster management| 82 disaster forecast models compared to the on-premise systems. fog computing attracts the cloud computing paradigm into the boundary of this network. fog computing is a distributed computing architecture, in which devices at the edge of the network are ubiquitously connected to offer storage services, communication, and flexible computation (anawar et al., 2018). due to the closeness to end-users in comparison with the cloud data centers, fog computing offers several advantages such as immediate notification services for realtime software, grade of service assurance and also low latency, location awareness (kotb et al., 2019). the center feature of this fog computing structure is that it provides data and compute analytics services more instantly and near the physiological apparatus which make this data, i.e. in the edge of this system, and so bypassing the wider internet (mahmood, 2018). the emergence of this internet of things (iot) has forced cloud computing to be united with fog computing, in order to prevent latency. evacuation wayfinding systems must not make use of the calculating capacity of the cloud and also apparatus but other individual devices in the vicinity to offer services (see fig. 2). edge computing and haze computing are very promising to support future wayfinding evacuation systems, because they not only utilize the computing power of portable and cloud devices, but also other individual devices in the vicinity to offer services (bi & gelenbe, 2019). fig. 2. fog computer architecture (zhu et al., 2013) 2. related work while exterior navigation is well recognized and accessible just about all weather conditions, indoor navigation nevertheless remains a technological discipline with no effective integrated solutions due to the increasing complexity of internal spaces and dynamic shift in certain specific partitions in large indoor locations, indoor navigation has become more important as it is helpful to aid individuals find their destination or evacuate from dangerous places. disasters could be either artificial or natural, and in the two situations, among the most significant elements in reducing loss of life is the time. the global positioning system (gps) can be used in the instance of outside navigation and is founded on satellite navigation by providing reliable location information on the consumer. once an evacuation occurs in confined spaces process can be more challenging since the motion and alternatives of evacuees are restricted from the surrounding atmosphere. emergency evacuation methods have 83 | international journal of informatics information system and computer engineering 1 (2020) 79 90 experienced a few phases: in the original human systems to the cloud-based wayfinding systems, which are still in their infancy wayfinding systems, to the currently booming in wireless sensor network-based (bi & gelenbe, 2019). disaster management system and accident detection may face challenges concerning latency and bandwidth awarded the centralized nature of the cloud service. an emerging concept that can help address these problems is fog computing that provides the promise of latency assistance, scalability and greater resilience (dar et al., 2019). the primary notion of fog computing would be to utilize the processing and storage resources at the edge of the network by deploying solutions on advantage devices that are open to decrease latency. the enhancement of complexity for indoor structures may cause many issues regarding to the evacuation system in case is indoor disaster. buildings are turning into a significant topic of recent disasters leading to huge quantities of casualty and damage costs for both the public and crisis managers. greater urbanization and sophistication of modern structures subtract emergency response processes and decision making particularly in indoor crisis scenarios where a lot of the data remains unrevealed to crisis managers before inputting the scenes exposing danger to individuals' lives as well as resources. additionally, unfamiliarity with all the indoor surroundings, restricted visualization because of smoke, and dropped and obstructed regions all raise the problem of crisis response and rescue in addition to undesirable drifting and doubt routing in indoor environments (tashakkori et al., 2015). (nikoohemat et al., 2020) are implementing a complete workflow, allowing the development of 3d models from the building point cloud and the extraction from these models of fine seed indoor networks to enable comprehensive emergency management route planning and navigation of the different types of agents. the method removes structural elements such as walls, tiles, ceilings and openings and reconstructs their volumetric forms. indoor emergency response plays an important role in disaster management, especially for a fire in a building, which must consider the evacuation of people in an environment that is indoor that is dynamic. a process of cooperation and coordination is necessary to include various actors and stakeholders in the handling of disasters. disaster management is the organization and alignment of all appropriate measures for creating, sustaining and strengthening preparation for, defending from, adapting to and recovering from humanled disasters. the topic is a multidisciplinary field of research where ict, in particular the internet of things, has a significant impact on the sensing and dynamic calculations of responses which increase or prevent the worst effects of disasters (bi and gelenbe, 2019). one of the significant problems for indoor emergency management is that the current indoor spatial structures do not fulfil emergency response criteria (tashakkori, 2017). the indoor search and rescue challenge can, therefore, not be easily used for decision taking in this area. one among the most important factors in indoor emergency response operations is awareness, which is created only after assessing all information that is available (dilo & zlatanova, 2011). the great deal of multi-domain sensory data generated asep id hadiana. fog computing architecture for indoor disaster management| 84 by increasingly advanced sensor networks that increase the burden of crisis navigation methods requiring time-critical response to the collection of information. interpretation and transmission traditional wsn-based emergency response systems, including functionality-identical detectors, have trouble providing optimum evacuation in a timely manner in the presence of all time-varying hazards due to this limitation of processing space, battery power and speed of communication (gelenbe & bi, 2014). fog computing has been widely researched in various fields, for example, the health sector, smart city, commercial industry, and other areas (hernándeznieves et al., 2020). the work discussed in (al-khafajiy et al., 2019) focusses on the development of a smart healthcare monitoring device that can remotely track the elderly. the technology mentioned in this paper is aimed at monitoring the physiological data of a person in order to identify particular conditions that could help to improve early intervention practices. the accuracy and analysis of sensory data acquired while transmitting disorder detection to a suitable career are achieving this. meanwhile (yan and su, 2016) has introduced an infrastructure to enhance current smart meter data storage and processing. the smart home, the building etc. is the user layer in this case. the fog layer consists of intelligent meters serving as fog nodes. they act as a particular data node that is considered a master node with modules that store the file name and storage location metadata. include also modules which duplicate and divide the collected data and then distribute it at fixed intervals to data nodes. using the mqtt protocol, (mekki et al., 2019) propose a remotely controlled ips. mqtt is an internet-of-things (iot) data exchange protocol widely found in the tcp / ip architecture. because of low bandwidth and the use by limitedresources devices, it is more suitable than other protocols, such as http. (dela cruz et al., 2016) has developed and designed the corresponding specification for wireless mesh network points using accessible hardware. apps have been introduced, such as message caching. the network, especially the mqtt messaging, has also been tested for reliability. 3. methodology we adopt the research methodology scheme described in fig. 3 following the indications in (kitchenham, 2004). fig. 3. the research methodology first, we provide a temporal interpretation of the literature in order to demonstrate the temporal conduct of research and the common interest in the fog computing paradigm qualitatively. secondly, we address in greater depth the fog computing model, which illustrates the need to incorporate them into iot. third, we examine the model of fog computing and the implementation scenarios together to assess the hot topics 85 | international journal of informatics information system and computer engineering 1 (2020) 79 90 and associated research concerns. fourth, we propose indoor disaster architecture based on fog computing. ultimately, we build on open issues and possible guidance about fog computing in the area of indoor disasters. 4. results and discussion the nist (national institute of standards and technology) fog computing model defines fog computing as a vertically model which enables distributed applications and services to be deployed, ensures low latency, reliable operation and eliminates the need for persistent cloud connectivity (battistoni et al., 2019). in fig. 4, we can see how the fog computing layer is structured. fig. 4. fog computing architecture (battistoni et al., 2019). from fig. 4, we can see that the data sources originating from end devices are collected and processed by fog computing. if necessary, it will be passed to cloud computing. the number of layers of fog computing can be more than one. with regard to efficiency and storage capacities, the fog computing processing and storage modules called fog nodes are heterogeneous. the cisco sets the standard terms for each level and outlines each level and level interaction and functionality, including the iot reference model (cha et al., 2018). the model of reference consists of seven levels. each level does not limit or limit the size of the respective component. for example, table 1 identifies cisco 's proposed reference iot model and related levels. table 1. cicco’s internet of things reference model. layer name detailed role 1 physical devices & controllers the “things” in iot 2 connectivity communication, processing units 3 fog computing data element analysis, transformation 4 data accumulation storage 5 data abstraction aggregation, access 6 application reporting, analytics, control 7 collaboration & processes involving people, business processes asep id hadiana. fog computing architecture for indoor disaster management| 86 according to (xu et al., 2016), the fog node serves the following purposes: • behaves as actual broker for mqtt clients, • serves as a platform for performing analytics at the fog node. • have to connect with the server to store the end-host broker and complete delayed review. the fact should be noted that an external fog node communication is used in a distributed broker environment, for example, to link to another fog node. messaging protocol is one of the important things in iot solutions. with regard to the question of which iot solutions have the messages protocol used, the results from this analysis have shown that the most used and adopted protocols are mqtt and http (dizdarević et al., 2019). mqtt and http rest are currently more mature and robust iot standards than other protocols. with several iot developers in its iot, fog or cloud deployments, mqtt and http, are the protocols of convenience. the publish / subscribe concept is used in this protocol. contrary to http using the request/response concept. pubsub (publish and subscribe) is an event-driven server that allows messages to be sent to the client when necessary. the contact center is at mqtt broker, which sends all messages, including the distribution channel. all customers, including the subject in the messages, who send a message to the broker. the subject is included in the broker's routing information. clients who want to receive messages may subscribe to a given topic, and the broker sends to the correct client all of the messages corresponding to the pattern of the subject. this method does not necessarily mean that clients need to know each other, but may easily interact using themes. mqtt is a lightweight coordination protocol for mq transportation. a broker can be connected to multiple customers or nodes in the mqtt network. the clients advertise certain data under a specific subject as a message. any other user who would like to receive these data subscribes to the related topic. the exchange of all data by the broker takes place. speaking solutions of this type are flexible because they need to be kept out of the ties between device producers and data consumers. in this indoor disaster architecture, each building is assumed to have smoke sensors and fire sensors, which will send data about fire events to mqtt broker server for later data processing as we can see in fig. 5. fig. 5. mqtt publisher-broker architecture in fig. 6, we can see that a disaster management system, which includes a fire brigade, police and medical service, acts as a mqtt subscriber. a disaster management system will get data or a warning from mqtt broker regarding the danger of a fire that occurs in a building. mqtt has been developed to 87 | international journal of informatics information system and computer engineering 1 (2020) 79 90 address the complexities of linking the increasingly physical world of sensors and actuators with a platform for information processing. reactions times, output, less battery use and lower bandwidth are key design criteria in the mobile environment. the mqtt protocol provides an advanced iot data control management and collection system. each control packet helps to reduce the data transmitted across the network for a smaller footprint and better battery life, and each bit of the packet. the mqtt broker is the main mediator between publishers and subscribers. this ensures that subscribers and publishers share messages through the broker node (see fig. 7). fig. 6. mqtt subscriber-broker architecture fig. 7. mqtt connection of the iot, fog and the cloud as in fig. 7, a raspberry pi is a mqtt client that is completely compatible with the mosquitto broker by installing the library of the mqtt paho. at the other hand, the broker refers to the higher abstraction layer of a fog machine node with more significant processing and storage capabilities. raspberry pi is connected to the temperature, smoke and fire sensors and releases data on the sensors to a broker fog node in the proposed intelligent scenario. in this case, the local server with the installed mosquitto library has a position for the broker, a simple personal computer outside the shelf. the information from local server/mqtt broker then transmitted to the cloud released to the local broker. here, the mqtt fog broker is used to connect all data to another cloud-based instance mqtt broker. related work so far has more commonly considered mqtt from the iot device layer to fog nodes. can be found in (dizdarević et al., 2019). asep id hadiana. fog computing architecture for indoor disaster management| 88 5. conclusion most people spend indoor time. the interior is more complicated than the exterior. in addition, today's building structures are becoming sophisticated and complex, so that disasters in the room can lead to problems. fire is one of the catastrophes in a building that often happens. they ought to handle accidents to reduce the likelihood of casualties. the internet of things is a current technological trend. via the internet of things, various developments exist. in various studies, disaster management using the internet of things and cloud computing was widely discussed. though, there are many issues with crisis recovery where the time and resources required are important, relies on cloud computing. of this function is fog computing. fog computing is used in this work as the foundation of the buildings' emergency response system. mqtt acts as a communication protocol with flexibility and speed advantages. this study proposes an indoor disaster architecture, mainly fire. this research is expected to be the basis for further research by utilizing all the advantages offered by fog computing in the context of handling a disaster. along with advances in artificial intelligence and deep learning technology, it is very open to research that can integrate fog computing with artificial intelligence and deep learning in supporting the indoor disaster management process. references al-khafajiy, m., baker, t., chalmers, c., asim, m., kolivand, h., fahim, m., & waraich, a. (2019). remote health monitoring of elderly through wearable sensors. multimedia tools and applications, 78(17), 24681-24706. anawar, m. r., wang, s., azam zia, m., jadoon, a. k., akram, u., & raza, s. (2018). fog computing: an overview of big iot data analytics. wireless communications and mobile computing, 2018. bastos, a., & vieira, v. (2015, may). indoor location systems in emergency scenariosa survey. in anais principais do xi simpósio brasileiro de sistemas de informação, 251-258. battistoni, p., sebillo, m., & vitiello, g. (2019, june). experimenting with a fogcomputing architecture for indoor navigation. in 2019 fourth international conference on fog and mobile edge computing (fmec), 161-165. bi, h., & gelenbe, e. (2019). a survey of algorithms and systems for evacuating people in confined spaces. electronics, 8(6), 711. bi, h., & gelenbe, e. (2019). emergency management systems and algorithms: a comprehensive survey. arxiv preprint arxiv:1907.04136. cha, h. j., yang, h. k., & song, y. j. (2018). a study on the design of fog computing architecture using sensor networks. sensors, 18(11), 3633. 89 | international journal of informatics information system and computer engineering 1 (2020) 79 90 dar, b. k., shah, m. a., islam, s. u., maple, c., mussadiq, s., & khan, s. (2019). delayaware accident detection and response system using fog computing. ieee access, 7, 70975-70985. dela cruz, a. a., parabuac, m. l. a., & tiglao, n. m. c. (2016, october). design and implementation of a low-cost and reliable wireless mesh network for firstresponse communications. in 2016 ieee global humanitarian technology conference (ghtc), 40-46. desmet, a., & gelenbe, e. (2013, october). reactive and proactive congestion management for emergency building evacuation. in 38th annual ieee conference on local computer networks, 727-730. dilo, a., & zlatanova, s. (2011). a data model for operational and situational information in emergency response. applied geomatics, 3(4), 207-218. dizdarević, j., carpio, f., jukan, a., & masip-bruin, x. (2019). a survey of communication protocols for internet of things and related challenges of fog and cloud computing integration. acm computing surveys (csur), 51(6), 1-29. gelenbe, e., & bi, h. (2014). emergency navigation without an infrastructure. sensors, 14(8), 15142-15162. hernández-nieves, e., hernández, g., gil-gonzález, a. b., rodríguez-gonzález, s., & corchado, j. m. (2020). fog computing architecture for personalized recommendation of banking products. expert systems with applications, 140, 112900. khan, a. a., yao, z., & kolbe, t. h. (2015). context aware indoor route planning using semantic 3d building models with cloud computing. in 3d geoinformation science, 175-192. kitchenham, b. (2004). procedures for performing systematic reviews. keele, uk, keele university, 33(2004), 1-26. klepeis, n. e., nelson, w. c., ott, w. r., robinson, j. p., tsang, a. m., switzer, p., ... & engelmann, w. h. (2001). the national human activity pattern survey (nhaps): a resource for assessing exposure to environmental pollutants. journal of exposure science & environmental epidemiology, 11(3), 231252. kotb, y., al ridhawi, i., aloqaily, m., baker, t., jararweh, y., & tawfik, h. (2019). cloud-based multi-agent cooperation for iot devices using workflownets. journal of grid computing, 17(4), 625-650. mahmood, z. (ed.). (2018). fog computing: concepts, frameworks and technologies. springer. asep id hadiana. fog computing architecture for indoor disaster management| 90 mekki, k., bajic, e., & meyer, f. (2019, april). indoor positioning system for iot device based on ble technology and mqtt protocol. in 2019 ieee 5th world forum on internet of things (wf-iot), 787-792. nikoohemat, s., diakité, a. a., zlatanova, s., & vosselman, g. (2020). indoor 3d reconstruction from point clouds for optimal routing in complex buildings to support disaster management. automation in construction, 113, 103109. palmieri, f., ficco, m., pardi, s., & castiglione, a. (2016). a cloud-based architecture for emergency management and first responders localization in smart city environments. computers & electrical engineering, 56, 810-830. ray, p. p., mukherjee, m., & shu, l. (2017). internet of things for disaster management: state-of-the-art and prospects. ieee access, 5, 18818-18835. tashakkori hashemi, s. h. (2017). indoor search and rescue using a 3d indoor emergency spatial model (doctoral dissertation). tashakkori, h., rajabifard, a., kalantari, m., & aleksandrov, m. (2015, november). indoor incident situation awareness using a 3d indoor/outdoor spatial city model. in 2015 2nd international conference on information and communication technologies for disaster management (ict-dm), 240-245. ujjwal, k. c., garg, s., hilton, j., aryal, j., & forbes-smith, n. (2019). cloud computing in natural hazard modeling systems: current research trends and future directions. international journal of disaster risk reduction, 38, 101188. xu, y., mahendran, v., & radhakrishnan, s. (2016, january). towards sdn-based fog computing: mqtt broker virtualization for effective and reliable delivery. in 2016 8th international conference on communication systems and networks (comsnets), 1-6. yan, y., & su, w. (2016, may). a fog computing solution for advanced metering infrastructure. in 2016 ieee/pes transmission and distribution conference and exposition (t&d), 1-4. zafari, f., gkelias, a., & leung, k. k. (2019). a survey of indoor localization systems and technologies. ieee communications surveys & tutorials, 21(3), 2568-2599. zhu, j., chan, d. s., prabhu, m. s., natarajan, p., hu, h., & bonomi, f. (2013, march). improving web sites performance using edge servers in fog computing architecture. in 2013 ieee seventh international symposium on service-oriented system engineering, 320-323. 103 | international journal of informatics information system and computer engineering 2(1) (2021) 103-112 utilization of internet of things on food supply chains in food industry hanhan maulana*, selvia lorena br ginting**, pramanda aryan***, muhamad restu fadillah****, rubi nurajmi kamal***** *japan advanced institute of science and technology **center for artificial intelligence technology, universitas kebangsaan malaysia, malaysia ***,****,*****departemen sistem informasi, universitas komputer indonesia, indonesia e-mail: * hanhan@email.unikom.ac.id a b s t r a c t s a r t i c l e i n f o this study aims to analyze the use of the internet of things (iot) in supporting the management of food supply chains (fscs) in the food industry. this research used qualitative research methods. the results obtained from this study are increasing the effectiveness and efficiency of the existing food supply chain in the food industry by applying the iot concept to food supply chain management. these results can be obtained because the iot concept is supported by various systems and technologies that can be implemented and developed so that iot can help identify and deal with existing problems more quickly while being able to assist in the decision-making process with information obtained through iot technology so that it will support development food supply chain management in the food industry. this study was conducted to see how much influence the internet of things (iot) has on food supply chain management in the food industry. article history: ___________________ keywords: internet, technology, supply, industry 1. introduction at this time, technology is developing so fast. one form of technological development is the use and access of the internet, which is quite extensive in various fields. the enlarger of technology, especially the internet, is shown by the emergence of many concepts in its use. one of the concepts of using the internet is the internet of things (iot). iot is a system of interconnected devices, using sensors and internet connections to collect various data and information (jones et.al., 2018). then the main key for the use of the internet of international journal of informatics, information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 2(1) (2021) 103-112 received 18 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 hanhan et al. utilization of internet of things on food supply chains in food industry...| 104 things in an industrial field is the concept of technology. if the food industry wants to apply intelligent management and automation to aspects of the supply chain, it is important to build a system that has good efficiency for the food industry based on the internet of things (zhang et.al., 2018). the food industry is quite complex. in the food industry, there are several industrial activities known as food supply chains (fscs). food supply chains are a major thing in the food industry because fscs show the flow of the production process to distribution to consumers, which broadly includes the production, processing, distribution, retail, and consumer processes. the application of iot in the food supply chain (fsc) is one of the promising applications. because its applications cover precision agriculture to food production, processing, storage, distribution, and consumption, it presents a promising potential to overcome challenges especially, in terms of control. several studies have been researched on the internet of things for the food supply chain. a study about the internet of things solution for the food supply chain was conducted by xiaorung, et al. in the study, it is stated that iot will create safer, more efficient, and sustainable fscs (zhao et.al., 2015). this is in line with jagtap & rahimifard. they claimed that implementing iot in the food supply chain will help provide high-level information and increase awareness about resource consumption across all actors in the food supply chain (jagtap & rahimifard, 2017). a study about the impact of the internet of things on the supply chain was conducted by mostafa, et al. in the study, it is stated that iot is one of the important technologies in facing industry 4.0. iot is considered as one of the greatest technology in improving performance and controlling capability in supply chain management (mostofa et.al., 2019). moreover, there is a study that shows the roles of food supply chain quality management by ben-daya, et al (ben-daya et.al., 2020). there is also another study that virtualizes food supply chains with the internet of things by verdouw, et al. in that study, it is stated that virtualization can play an important role in meeting the specific challenges of the food supply chain, including dealing with perishable products, unpredictable supply variations, and food products that have strict safety and requirements (verdouw et.al., 2016). therefore, from these studies, we can see the role of technology is quite important, especially in supply chain management. the purpose of this research is to identify, analyze, and depict the use or application of iot in the food industry, particularly in the context of food supply chain management. the making of this paper is done by collecting various data and then analyzing the data, which will produce a result. the results of this analysis show an overview related to the management system in the food supply chain, which will indirectly show an increase in the effectiveness and efficiency of the food supply chain management system in the food industry with the application of iot. 2. method the method used in this research is the descriptive quantitative method, the research method used is qualitative, namely research that is descriptive and tends to use analysis. this research 105 | international journal of informatics information system and computer engineering 2(1) (2021) 103-112 method aims to identify and describe the events experienced by the research subjects, such as perceptions, behavior, and actions, which describing in the form of words and language in a special context and using various methods carried out as a whole. the research was carried out in several steps, starting from data collection, data analysis that would produce a finding, which can be seen further in the image below. (see fig.1). fig. 1. method figure 1 shows that the author conducts research methods on "utilization of the internet of things on food supply chains in food industry" through: a. the data and information collection method used in this article is a literature study. this literature study is used to obtain data and information relevant to the existing problems. literature studies are carried out by reading, studying, reviewing literature in the form of journals, articles, and books on the internet related to topics and problems in research in articles. the theme raised in this article is related to the use of iot in the food industry, especially in food supply chain management. b. after obtaining the necessary data, data analysis was carried out, then the data were analyzed using the literature review method. this analytical method is used by systematically identifying the studies that have been obtained relating to the use of the internet of things for food supply chains, especially in the food industry. c. the results of this analysis can identify the utilization or implementation of the internet of things in food supply chain management in the food industry. 3. results and discussion the food supply chain generally runs in several stages, starting from production, processing, distribution, retail, and consumers. however, in the food industry, the food supply chain is more focused on production, processing, and distribution. a. production this stage of production is related to agriculture activities that exist on a farm. at the production or agriculture stage, several activities or operations in the agricultural sector can utilize iot technology, such as irrigation control and pest control. at the production stage, there are also livestock sectors, such as smart poultry farms, which also need to be considered regarding the use of iot in the food industry. 1. irrigation control irrigation is important in the agriculture stage. the net ratio of benefits to the entire agricultural sector water productivity in agriculture is defined. one of the keys to increasing water productivity in agriculture is to maximize irrigation scheduling, so agriculture can save water without affecting existing yields. the iot concept can be used to create smart irrigation. with smart hanhan et al. utilization of internet of things on food supply chains in food industry...| 106 irrigation, workers can monitor irrigation and can turn on pumps without having to visit the plant location. the iot concept can be done by utilizing an internet network for all sensors on agricultural land (soil and weather sensors). data related to .environment can be obtained and sent to workers through applications. a smart irrigation system can be seen further in the image below (see fig 2). fig. 2. smart irrigation system figure 2 shows the smart irrigation system. it can be seen the stages related to the smart irrigation system starting from receiving data on environmental temperature, soil, and temperature humidity by sensors at agricultural locations. then by utilizing an internet connection, the data will be sent to internet cloud services and then stored on internet cloud services. if there are abnormal conditions related to environmental temperature, soil, and humidity, the monitoring section or user will receive a report. based on the report, the user can give orders through existing applications to the irrigation control system so that irrigation can be done remotely. 2. pest control the reliability of crop disease and pest management systems depends on three aspects: sensing, evaluation, and treatment (supported by iot technologies) (zhang et.al., 2018). utilization of iot in pest management can be done by utilizing tools that have sensors to obtain data or information related to the environment, the state of plants needed to control pests. examples of tools used to control pests are iot automated traps, which have the function of automated camera traps and adjustable spray timing. the automated traps can be seen further in the image below (see fig.3). fig. 3. iot automated traps (semios) source : www.semios.com figure 3 shows one of the tools that utilize iot technology in controlling pests, namely automated traps. in the series of automated traps, there are automatic camera traps, spray timers, and variable rate mating disruption. automatic camera traps will send pictures as well as the number of catches related to pests daily. then the spray timer is related to setting the anti-pest spraying device, which can be adjusted as needed. then the variable-rate mating disruption is to predict the level of pheromone application in response to current pest behavior. pest control system through automated traps can be seen further in the image below (see fig. 4). 107 | international journal of informatics information system and computer engineering 2(1) (2021) 103-112 fig. 4. pest control system figure 4 shows an overview of a pest control system that uses automated traps. the data generated through iot automated traps will be transferred to the internet cloud system. the entire system that monitors agriculture environmental conditions by utilizing iot automated traps will automatically provide reports regarding pest problems and automatically provide workers with early warnings regarding pest problems. 3. smart poultry farm in general, the success of poultryrelated production depends on comfortable environmental conditions such as temperature, humidity, and lighting. the iot concept can be used to determine the condition of poultry through existing sensors to obtain data related to temperature, humidity, and lighting. then, the data will be directed to internet cloud services and stored. the monitoring department related to the condition of poultry will use available data to determine the real-time condition of poultry, such as stress levels and comfort based on calculations related to temperature, humidity, and lighting. when compared to traditional operations, poultry farming by utilizing iot has advantages in the form of better reliability and expansion in obtaining information so that it can assist in making decisions in the field in managing poultry, as well as increasing remote control capabilities that will increase the effectiveness and efficiency of operations in the field poultry farming. these advantages will lead to an increase in the quality of poultryrelated to health and welfare. b. processing this processing stage is related to the activity of converting or processing raw materials into a product. in the food industry, it is certainly related to food products. iot can collect data in realtime with this data, optimization of related processes in the food industry can be carried out. with real-time information gathering and intelligent algorithms, the control software can make better decisions to avoid deviations in actuator driving. with iot technology, monitoring capabilities in the processing section can be more easily carried out. monitoring at the processing stage with real-time data collection by iot has a positive impact on increasing maintenance. for hanhan et al. utilization of internet of things on food supply chains in food industry...| 108 example, the use of sensors to monitor the temperature on the machine can be used to take preventive measures when there is information related to temperatures that are out of range (exceeding normal limits) this will certainly prevent engine damage. iot technology also provides end-to-end transparency capabilities in near realtime. this relates to traceability, which at the processing stage refers to the ability to record or obtain data through rfid tags or other tracking media so that all product processing steps can be monitored better. with traceability, it is possible to optimize the entire line at the processing stage and will increase the efficiency of the processing stage in the food industry. for example, in sauce production, an sme using a machine on the processing stage that utilizes iot technology to improve traceability shows an increase in income of 30% from previous income (apriharta et.al., 2021). this example shows that iot technology can increase the effectiveness and efficiency of the processing stage. in the processing stage, it is also necessary to pay attention to the packaging of food products. food packaging is used for protecting food products from external influences such as temperature, lighting, humidity, and others. technological developments such as the emergence of iot are changing packaging to be more interactive and safe. smart packaging with interactive and intelligent packaging concepts is based on the interaction between the packaging environment and food to provide active protection for food products. smart packaging utilizes sensors to monitor food quality and safety. monitoring the quality and safety of food products can be done online with smart packaging combined with the iot concept. the concept of iot in smart food packaging can be seen further in the image below. (see fig. 5). fig. 5. smart food packaging system 109 | international journal of informatics information system and computer engineering 2(1) (2021) 103-112 figure 5 shows a smart food packaging system that utilizes the iot concept. by embedding sensors on food product packaging such as rfid or smart labels, then utilizes an internet connection, data such as temperature, humidity, and lighting can be directed to the internet cloud service. these data can be used to determine the condition of food, especially related to the quality and safety of food products. then based on these data, an analysis will be carried out, which will influence the decisionmaking regarding food product packaging. the utilization of iot has good potential in developing new sensor systems embedded in food product packaging so it can improve the ability to convey information related to weight and volume control of food products. c. distribution this distribution stage is related to warehousing activities and the distribution of processed products to retailers. 1. warehouse management warehouse management is an important thing in the food industry because the warehouse is a place to store food products that have been processed. the warehouse can store so many products, so it must be managed properly and optimally to ensure fast and precise performance in all functions so that customer demands can be met while ensuring the condition of the food products in the warehouse. warehouse management must know all the products to be distributed, using iot technology to help warehouse management identify existing food products. warehouse management must also know the conditions related to food products in the warehouse, using iot technology to help warehouse management receive information when abnormal conditions occur in the warehouse, such as inappropriate temperatures, inappropriate product locations, and low product stocks. in warehouse management, to ensure good performance in all warehouse functions, generally using sensor technology, rfid, and perceptual technologies such as voice and video monitoring (juntao, 2016). improving traceability is one of the keys to improving warehouse management capabilities because, with good traceability, the ability to monitor and identify products in the warehouse will increase. the warehouse management system can be seen further in the image below (see fig. 6). figure 6 shows a warehouse management system by utilizing iot. the iot concept on warehouse management system can be utilized to improve traceability is to use rfid tags as a tool to identify, track and obtain data from a target object in the form of food products, then wireless sensor network and actuators to monitor related data obtained by rfid and direct the data obtained to internet cloud services through a network gateway. internet cloud services will be a place to store data related to food products in the warehouse. hanhan et al. utilization of internet of things on food supply chains in food industry...| 110 fig. 6. warehouse management system 2. product distribution the distribution of food products needs to pay attention to the quality, freshness of the product, and damage to the product. if the distribution of products does not carry out strict controls, then food products are susceptible to bacteria, accelerate the rate of spoilage, affect product quality and threaten consumer health. during the distribution process, the temperature and humidity of the product must be within a certain range. otherwise, customer satisfaction with the distributed product will be affected (chen et.al., 2020). the selection of distribution channels is also important because to distribute food products, time is crucial, especially to ensure the quality of food products. key technologies related to the use of iot in the distribution of food products are rfid, sensor technology, video surveillance, and cellular communication technology in the form of the global position system (gps) (wen & lv, 2019). the iot concept can be utilized in product distribution by using humidity and temperature sensors on trucks transporting food products. the humidity and temperature sensors will be combined with rfid embedded in the product packaging so that productrelated information can be recorded. with an internet connection, the information contained in the rfid will be directed to the monitoring center so that through the monitoring center, the information related to humidity and temperature can be monitored in realtime. with gps installed on the product transport truck, information related to the vehicle location, trip status, information related to the lanes can be passed or can be obtained. so that the monitoring center can find out the 111 | international journal of informatics information system and computer engineering 2(1) (2021) 103-112 condition of the transport truck and ensure on-time delivery. 4. conclusion this paper studies the food supply chain in the food industry and analyzes the situation according to real conditions. in general, food supply chain management in the food industry is still running manually, thus creating ineffectiveness and inefficiency in the course of the food supply chain, especially in obtaining information. so, in this paper, we introduce the internet of things that can be utilized in the food supply chain in the food industry and the internet of things-based information delivery system. this system applies various technologies combined with the concept of the internet of things so that the distribution of information in the food supply chain can be carried out more effectively and efficiently. references apriharta, taufani, a. r., kusumaningrum, i. k., maharani, s. n., & firmansyah, a. (2021). penerapan iot pada proses produksi saos dengan sistem less-contact didukung database-smart app untuk ukm sejahtera sentosa. jurnal abdimas berdaya, 4, 50-61. ben-daya, m., hassini, e., bahroun, z., & banimfreg, b. h. (2020). the role of internet of things in food supply chain quality management: a review. quality management journal. american society for quality. chen, j., xu, s., chen, h., zhao, c., & xue, k. (2020). research on optimization of food cold chain logistics distribution route based on internet of things. in journal of physics: conference series, 1544. jagtap, s., & rahimifard, s. (2017). utilization of internet of things to improve resource efficiency of food supply chains. in ceur workshop proceedings, 2030, 8–19. jones, n.b., graham, c.m., 2018. can the iot help small businesses? bulletin of science, technology and society, 38, 3–12. juntao, l. (2016). research on internet of things technology application status in the warehouse operation. international journal of science, technology and society, 4(4), 63. mostafa, n., hamdy, w., & alawady, h. (2019). impacts of internet of things on supply chains: a framework for warehousing. social sciences, 8(3). verdouw, c. n., wolfert, j., beulens, a. j. m., & rialland, a. (2016). virtualization of food supply chains with the internet of things. journal of food engineering, 176, 128–136. wei, j., & lv, s. (2019). research on the distribution system of agricultural products cold chain logistics based on internet of things. in iop conference series: earth and environmental science, 237. hanhan et al. utilization of internet of things on food supply chains in food industry...| 112 zhang, j., qu, x., & sangaiah, a. k. (2018). a study of green development mode and total factor productivity of the food industry based on the industrial internet of things. ieee communications magazine, 56(5), 72–78. zhang, l., dabipi, i. k., & brown, w. l. (2018). internet of things applications for agriculture. in internet of things a to z: technologies and applications, 507–528. zhao, x., fan, h., zhu, h., fu, z., & fu, h. (2015). the design of the internet of things solution for food supply chain. in proceedings of the 2015. international conference on education, management, information and medicine, 8. 23 | international journal of informatics information system and computer engineering 2(2) (2021) 23-30 sustainable and digitized certification of palm oil production : its impact on the environment in indonesia angga fauzan pratama wageningen university and research e-mail: anggafp@gmail.com a b s t r a c t s a r t i c l e i n f o the purpose of this study was to explain sustainable and digitized certification of palm oil production and its impact on the environment in indonesia. in indonesia, palm oil producers consist of private plantations as the largest national producers (54%), smallholders (39%), and state-owned plantations (7%). the management of smallholder oil palm still have limitations to accessing technology, production facilities, institutions, and marketing. that is the reason all the palm oil companies need certification to make them standard in facing environtment problems. with the system of digitizing certificate, especially for smallholder companies, it enable to help them transformed the business processes as well as waste treatment. article history: received 18 nov 2021 revised 20 nov 2021 accepted 25 nov 2021 available online 26 dec 20218 ___________________ keywords: sustainable and digitized, palm oil, production, impact, environment 1. introduction deforestation gives an adverse effect on the environment, including climate change. forest cover conversion releases co2 into the atmosphere, which worsens the greenhouse effect. besides altering the carbon level in the atmosphere, deforestation also leads to a change in surface temperature. in the subtropics, deforestation leads to a cooler temperature, while in the tropics, deforestation leads to a warmer temperature (longobardi et al., 2016). deforestation was found to be affecting the moisture cycle as well, and therefore, affecting regional precipitation (chambers and artaxo, 2017). in indonesia, palm oil plantation becomes the dominant drivers of deforestation. increasing global palm oil demand compels stakeholders to clear more forest for more expansive plantation. between 2001-2016, two million hectares of forest (26% of total deforestation) has been cleared for palm oil plantation (austin et al., 2019). at the same time, palm oil contributes positively international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(2) (2021) 23-30 pratama et al. sustainable and digitized certification of palm oil… | 24 to the indonesian economy. palm oil is the main export commodity of indonesia agriculture exports and has become indonesia's strategic industry. palm oil plantation also contributed to the development of the rural community (rist et al., 2010). study shows that palm oil farmer in the rural is more prosper than farmers of other commodities (kubitza et al., 2018). these positive impacts on the economy and development hinder the government of indonesia (goi) from stopping palm oil expansion. the forest cover loss by palm oil expansion has been noticed by international communities and leads to the discouragement of using palm oil. several standards and certifications for sustainable palm oil (spo) were introduced to prevent further discouragement. the roundtable of sustainable palm oil (rspo) is the most prominent spo certification body. rspo established the first sustainability standards for palm oil production and has certified over two million hectares of palm oil plantation in indonesia (rspo, 2020). the management of smallholder oil palm still have limitations to accessing technology, production facilities, institutions, and marketing. that is the reason all the palm oil companies need certification to make them standard in facing environtment problems. with the system of digitizing certificate, especially for smallholder companies, it enable to help them transformed the business processes as well as waste treatment. the digitization of sustainable smallholder oil palm plantations is designed to transform the business processes carried out by oil palm farmers who join cooperatives into a form of digital work patterns while still providing roles and appreciation for their permanent work from farmers, by farmers, and for farmers, so that smallholder oil palm plantations are industrial-based agricultural activities. era 4.0 [2]. the rspo certification is derived from the sustainable development pillars, which are the environmental pillar, economic pillar, and social pillar. this review will solely focus on the environmental pillar of rspo certification. in rspo certification, the environmental sustainability concern is represented in the seventh principle, which is to protect, conserve and enhance ecosystems and the environment (rspo, 2020). this principle is implemented in several criteria for the production process, which includes greenhouse gas (ghg) emission reduction, forest fire ban, and deforestation prohibition (rspo, 2020) palm producing countries, including indonesia, try to improve palm oil image through supporting and implementing spo certification, including rspo. despite the certification effort, palm oil is still viewed negatively. in 2018, under renewable energy directive, the european union (eu) members agreed to stop using palm oil for biofuel by 2030 due to palm oil's link to deforestation. the sustainable effort by goi that has been conducted through standards and certification seems to be neglected or considered to be not enough by the eu. the author interested in writing a literature review related to the arising issue and raises a research question: "how has and how can sustainable palm oil certification impact the environment of indonesia?" this literature review will discuss five recent literature regarding spo and its impacts, which are by (cattau et al.,2016), (carlson et al., 2018), (morgans et al., 2018), (gatti et al., 2019), and schmidt and (de rosa., 2020). this literature review 25 | international journal of informatics information system and computer engineering 2(2) (2021) 23-30 has the purpose of understanding the environmental impacts of spo certification from different researches. this literature review will be divided into four sections. the first section is the introduction and background of the review topics. the methodology used by each literature will be discussed in the second section. the results of each literature will be discussed in the third section, and the last section will discuss the conclusion of this literature review. 2. methodology this section will discuss the data and method that was used by the authors in assessing environmental impacts of spo certification. before comparing the data and method used by each authors, it is worth noting that each authors have different objectives. research by (cattau et al., 2016) is trying to reveal the impacts of rspo in reducing the number of forest fire through 2012-2015 on the island of sumatera and kalimantan, indonesia. (carlson et al., 2018) study covers broader aspects and wider timespan that (cattau et al., 2016). (carlson et al., 2018) investigate the effect of rspo to forest fire and deforestation in indonesia from 2001 until 2015. (morgans et al., 2018) study covers even broader than (calson et al., 2018) by trying to assess rspo impacts on all sustainability objectives. (gatti et al., 2019) are investigating the deforestation on spo certified plantation in indonesia between 2001-2016. meanwhile, (schmidt and de rosa, 2020) are comparing greenhouse gas emission from rspo and non-rspo plantation. from these differences in objectives, different data and method are used in the respective research. in general, the methods that were used by the authors can be categorised into spatial analysis and modelling. four out of five papers in this review, which are (cattau et al., 2016), (carlson et al., 2018), (morgans et al., 2018), and (gatti et al., 2019), use spatial analysis. only (schmidt and de rosa, 2020) do not use spatial analysis in their research. (cattau et al., 2016) are using indonesia's oil palm concession map obtained from global fire watch (gfw) produced by indonesia's ministry of forestry. (cattau et al., 2016) also use data from greenpeace that provides companies' total oil palm concession and total rspo certified oil palm concession. (cattau et al., 2016) trying to accommodate the intentional use of fire in oil palm planting and clearing by inspecting through google earth imagery to identify oil palm planting and clearing phase, and only compares forest fire in the developed period. (cattau et al., 2016) use fire detections data from the moderate resolution imaging spectroradiometer (modis) active fire detections by nasa firms. (carlson et al., 2018) digitise map from rspo secretariate and website to obtain spo certified area. (carlson et al., 2018) also use data from greenpeace and sawit watch, which are produced by goi. (carlson et al., 2018) combine the data and removed smallholders' plantation from the data so that only industrial plantation is being assessed. (carlson et al., 2018) use modis data, same as (cattau et al., 2016), to identify fire occurrences. meanwhile, to detect deforestation, (carlson et al., 2018) are using shuttle radar topography mission (srtm). (carlson et al., 2018) employ a combination of propensity score matching technique and panel methods to determine the impacts of spo certification. pratama et al. sustainable and digitized certification of palm oil… | 26 (morgans et al., 2018) use palm oil concession map in 2014 from the world resources institute (wri) made by goi. (morgans et al., 2018) also use data from rspo annual communication of progress (acop), sustainable palm oil transparency toolkit (zoological society of london sustainable palm oil platform) and global forest watch (gfw) to support and verify the wri's data. (morgans et al., 2018) use two metrics in assessing environmental sustainability impacts of spo certification, which are (1) presence and density of orangutan, and (2) number of fire hotspots detected. morgans et al. (2018) also use a propensity score matching technique to determine impacts of spo certification. (gatti et al., 2019) use forest cover change data from (hansen et al., 2013) and from gfw (accessed in 2018). (gatti et al., 2019) also use palm oil production and concession data from goi, rspo, and greenpeace. from these data, (gatti et al., 2019) determine the number of forests lost in indonesia by calculating the area of forest cover lost that overlaps concession area. (schmidt and de rosa, 2020) apply the life cycle assessment (lca) to compare the total greenhouse gas emission produced by spo certified palm oil and non-certified palm oil. life cycle assessment (lca) calculates the total emissions of a product by summing the emissions produced at different phases of the production to consumption (mukarjee and sovacool, 2014) and it fit to assess the environmental performances of products or production systems (schmidt and de rosa, 2020). to be able to calculate the total emission, a background database is needed in lca. (schmidt and de rosa, 2020) uses the latest version of the global hybrid environmentally extended multi-regional input-output (eemrio) database on exiobase v3 as the background database for the model. meanwhile, for data in the production processes, (schmidt and de rosa, 2020) use various sources, such as rspo reports, statistical data, and modelling approaches. 3. result in general, the findings by the authors can be grouped into three categories, which are (1) spo certification impact on forest fire, (2) spo certification impact on deforestation, and (3) spo certification impact on ghg emissions. the impact of spo certification on forest fire was studied by (cattau et al., 2016), (carlson et al., 2018), and (morgans et al., 2018). the three authors have different findings. (morgans et al., 2018) have the most general approach in assessing spo certification impacts on the environment. (morgans et al., 2018) only distinguish spo-certified plantation and non-spocertified. it was found that there is no significant impact of spo certification on forest fire and forest fire increased on both categories between 1999-2015 (morgans et al., 2018). (carlson et al., 2018) find a different result by discovering that most fire activity in indonesia's forest occurs before spo certification initiated by goi. however, based on significance statistics, (carlson et al., 2018) also agree that spo certification has no impact on fire activity. the decrease in fire activity happens not only in spo-certified plantations but also in non-spo-certified plantations. the study by (cattau et al., 2016) is distinguishing the land types (peatlands and non-peatlands), weather (dry and wet years), and certification 27 | international journal of informatics information system and computer engineering 2(2) (2021) 23-30 status (spo-certified and non-spocertified). (cattau et al., 2016) are able to find significant difference reduction of fire activity on non-peatlands during wet years. however, in peatlands and nonpeatlands in a dry year, fire activity is not significantly different (cattau et al., 2016). spo certification impact on deforestation was covered by two studies in this review. (carlson et al., 2018) noted that deforestation has higher rate before spo certification initiative, although deforestation still continues even after spo certification implementation. forest area of 91 km2 was lost in certified plantation area with 24 km2 of peat and 23 km2 of primary forest (carlson et al., 2018). significance statistics shows that spo certification contributed to a 33% reduction of deforestation and prevents 21 km2 ± 2.8 km2 of forest loss through 2015 (carlson et al., 2018). (carlson et al., 2018) noticed a bias in spo certification pattern and mentioned that lower deforestation on spo certified plantation is mostly due to the less forest available. however, (carlson et al., 2018) are optimistic that certification on plantations with more forest can leads to more significant forest saving. (gatti et al., 2019) disapproves (carlson et al., 2018) statements and points the proportion of deforestation in certified concession compared to the total deforestation is still relatively high, and even in 2015, the deforestation in certified concession was higher than in noncertified concession. the spo certification is also not sustainable because the plantations that are currently sustainable related to high forest degradation in the past (gatti et al., 2019). in this review, the reduction of ghg in the palm oil production process through spo certification was solely studied by (schmidt and de rosa, 2020) .modelling by (schmidt and de rosa, 2020) finds that spo-certified production is significantly more environmental-friendly than noncertified production. spo-certified product has a lower carbon footprint than a non-spo-certified product. it was found that spo-certified product has 36% lower ghg emission than the non-spocertified product. higher yields, peatland and nature protection, and biogas capture potential in spo certification promotes the benefits of ghg emission reduction (schmidt and de rosa, 2020). 4. discussion the modelling by (schmidt and de rosa, 2020) proves that spo certification has the potential to reduce palm oil negative contribution to the environment. lower ghg emission can generate more sustainable palm oil production. however, in practice, spo certification is not convincing, which is depicted by the conflicting results of some studies. spo certification was found to decrease fire activity in a specific land type (nonpeatland) during a specific period (wet years) (cattau et al., 2016). however, forest fire reduction due to spo certification cannot found on peatlands and dry years (cattau et al., 2016). studies that do not classify forest and see the forest as a whole also cannot find any significant impact of spo certification on forest fire incidence (carlson et al., 2018; morgans et al., 2018). the comparison of these three studies shows that the scope of study affects the observed result. detailing the scope of study will reveal a more accurate picture of the impacts of spo certification. concerning the spo certification policy, indicates that spo certification only positively impacts in a particular condition and have not been able to make a broad impact. pratama et al. sustainable and digitized certification of palm oil… | 28 difference in opinion can also be found on the claim of spo certification impacts to deforestation rate. (carlson et al., 2018) state that spo certification able to decrease deforestation rate and can have a more significant impact if it is implemented more widely (carlson et al., 2018), but (gatti et al., 2019) challenge the claim and states no significant impact of spo certification on deforestation rate. in this case, a difference in perspective is the cause of different opinion. in addressing the environmental impacts of palm oil plantation, (gatti et al., 2019) also consider the preceding land cover in the existing plantation. gatti et al. (2019) regard the forest clearance that was required to open the spo-certified plantations as deforestation. on the other hand, (carlson et al., 2018) only count the deforestation that occurs in the existing plantation and disregard the previously existing forest in the plantation area. 5. consclusion spo certification is initiated as an action to reduce the negative impacts of oil palm plantations. if spo certification is enacted properly, ghg emission of indonesia will be reduced. however, in practice, the implementation of spo certification in indonesia has not brought a convincing result. the positive impact of spo certification on reducing forest fire and deforestation is still limited and not convincing. more studies are required to discover the source of ineffectiveness and improve the effectiveness of the policy. deforestation and forest fire in palm oil plantation are still occurring in indonesia, and we still need to work against nature destruction and degradation. although so far spo certification implementation in indonesia have not brought a convincing result, a study has proven that spo certification has the potential to bring significant benefits. the stakeholders (governments, industries, and societies) need to work together to maximise the potential benefits of spo certification that are yet to be achieved references austin, kemen g., et al. "what causes deforestation in indonesia?." environmental research letters 14.2 (2019): 024007. carlson, kimberly m., et al. "effect of oil palm sustainability certification on deforestation and fire in indonesia." proceedings of the national academy of sciences 115.1 (2018): 121-126. carter, claire, et al. "palm oil markets and future supply." european journal of lipid science and technology 109.4 (2007): 307-314. 29 | international journal of informatics information system and computer engineering 2(2) (2021) 23-30 cattau, megan e., miriam e. marlier, and ruth defries. "effectiveness of roundtable on sustainable palm oil (rspo) for reducing fires on oil palm concessions in indonesia from 2012 to 2015." environmental research letters 11.10 (2016): 105007. chambers, jeffrey q., and paulo artaxo. "biosphere–atmosphere interactions: deforestation size influences rainfall." nature climate change 7.3 (2017): 175 176. gatti, roberto cazzolla, et al. "sustainable palm oil may not be so sustainable." science of the total environment 652 (2019): 48-51. goodland, robert. "the concept of environmental sustainability." annual review of ecology and systematics 26.1 (1995): 1-24. kubitza, christoph, et al. "the economics behind an ecological crisis: livelihood effects of oil palm expansion in sumatra, indonesia." human ecology 46.1 (2018): 107 116. longobardi, patrick, et al. "deforestation induced climate change: effects of spatial scale." plos one 11.4 (2016): e0153357. morgans, courtney l., et al. "evaluating the effectiveness of palm oil certification in delivering multiple sustainability objectives." environmental research letters 13.6 (2018): 064032. rist, lucy, laurene feintrenie, and patrice levang. "the livelihood impacts of oil palm: smallholders in indonesia." biodiversity and conservation 19.4 (2010): 1009-1024. rspo. "principles and criteria for the production of sustainable palm oil." roundtable on sustainable palm oil (2020), https://rspo.org/resources/certification/rspo-principles-criteriacertification. accessed 18 october 2020. pratama et al. sustainable and digitized certification of palm oil… | 30 schmidt, jannick, and michele de rosa. "certified palm oil reduces greenhouse gas emissions compared to non-certified." journal of cleaner production (2020): 124045. "impact." roundtable on sustainable palm oil. rspo, 2020. https://www.rspo.org/impact. accessed 18 october 2020. 1 | international journal of informatics information system and computer engineering 1 (2020) 91 102 dimensional speech emotion recognition from acoustic and text features using recurrent neural networks bagus tris atmaja1,2 reda elbarougy3, masato akagi2 1 sepuluh nopember institute of technology, surabaya, indonesia 2japan advanced institute of science of technology, nomi, japan 3damietta university, new damietta, egypt e-mail: bagus@ep.its.ac.id a b s t r a c t s a r t i c l e i n f o emotion can be inferred from tonal and verbal information, where both features can be extracted from speech. while most researchers conducted studies on categorical emotion recognition from a single modality, this research presents a dimensional emotion recognition combining acoustic and text features. a number of 31 acoustic features are extracted from speech, while word vector is used as text features. the initial result on single modality emotion recognition can be used as a cue to combine both features with improving the recognition result. the latter result shows that a combination of acoustic and text features decreases the error of dimensional emotion score prediction by about 5% from the acoustic system and 1% from the text system. this smallest error is achieved by combining the text system with long short-term memory (lstm) networks and acoustic systems with bidirectional lstm networks and concatenated both systems with dense networks. article history: received 17 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 lable online 09 sep 2018 ___________________ keywords: speech emotion, neural network, lstm. international journal of informatics, information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 1 (2020) 91 102 b t atmaja, et al. dimensional speech emotion recognition...| 92 1. introduction the demand for recognizing emotion in speech has grown increasingly as a human emotion can be expressed via speech, and many applications, such as call center, telephone communication, and voice messages, can benefit from this speech emotion recognition. the study of speech emotion recognition was established some decades ago using unsupervised learning and a small amount of data. advancements in computation hardware and in the development of larger speech corpus have enabled us to analyze emotion in a speech detecting emotion is useful to investigate whether a student is confused, engaged, or certain when interacting with a tutorial system or whether a caller to help a line is frustrating or not (jurafsky et al., 2014). by gaining knowledge of emotion from student and caller in both cases, proper action can be taken to avoid the worse condition. the degree of emotion (in the numeric score) in both cases is more relevant than the category of emotion (joy or sad, for example). these are two examples where dimensional emotion is more informative than categorical emotion. although research on emotion recognition has been conducted progressively, most research are focused on recognition of categorical emotion such as in (griol et al., 2019; chen et al., 2018; atmaja et al., 2019). as shown by the previous two examples, a dimensional approach of emotion recognition is more informative in such cases. recognizing the degree of emotion is a more challenging task as it tries to predict the numerical score rather than a category. this type of task is a class of logistic regression. research investigating dimensional emotion recognition in a text is reported by calvo et al., 2020. they found by using the same classifier, i.e., non-negative matrix factorization (nmf), both categorical and dimensional emotion recognition obtain a similar result. they used emotional terms from an affective dictionary as text features for the dimensional task. in speech emotion recognition, the study of dimensional emotion recognition is reported by (giannakopoulos et al., 2009) using a small dataset from videos, ten dimensions of acoustic features, and k-nearest neighbor (knn) to estimate emotion degree. the results indicate that the resulting architecture can estimate emotion states of speech from movies with sufficient accuracy (valence: 24%, arousal: 35%, in terms of r2 statistics). both dimensional text and speech emotion recognition above used a non-deep neural network (dnn) method due to the time and size of data. another challenge in speech emotion recognition, besides a dimensional approach, is the strategy for extracting features. the features are the input of an emotion recognition system, and the performance of the system depends on those features. an issue to be considered when extracting features for speech emotion recognition is the necessity of combining speech (acoustic feature) with other types of features (el ayadi et al., 2011). we choose text features as it can be extracted from speech via automatic speech recognition (asr). the combination of these acoustic and text features is expected to improve the performance of the emotion recognition rate compared to the use of single modality i.e., acoustic feature or text feature only. this paper presents a dimensional speech emotion recognition from a multimodal dataset. the purposes of this work are (1) to examine whether the fusion 93 | international journal of informatics information system and computer engineering 1 (2020) 91 102 of two related features can decrease the error of dimensional emotion recognition and (2) to find the best dnn architecture for a list of dnn layer combination. a deep learning-based classifier from the category of recurrent neural network has been built for this purpose. two types of features are used: acoustic and text features. for each feature, a set of networks is stacked. the two networks from acoustic and text features are then concatenated using late fusion architecture. the result shows that the proposed method can improve the performance compared to the method that used acoustic or text features only. the evaluation is presented in terms of mean squared error (mse), mean absolute error (mae), and mean absolute percentage error (mape). to extend this work, a discussion to evaluate the metric used in this research is summarized at the end of the paper. 2. dataset the iemocap (interactive emotional dyadic motion capture) database developed by the university of southern california was used in this research (busso et al., 2008). a number of 10039 turns (utterances) are recorded and measured, including included speech, visual, text, and motion capture (face, head, and hand movement). from those modalities, speech signal and text transcription are used. the dimension labels are given for valence, arousal, and dominance (vad) in a range of 1 to 5 via self-assessment manikins (sams). all utterances on this dataset are used in this research. from these data, 80% is used for training, and 20% is used for the test. twenty percent of the training data is used for validation. fig. 1. proposed dimensional speech emotion recognition from acoustic and text features. the dash line between label and dataset means that label is obtained from dataset directly. 3. proposed method a proposed method of this research paper can be split into two parts: feature extraction and dimensional emotion classifier. a block diagram of the proposed system is shown in fig. 1. from the dataset, two features are extracted: acoustic and text features. the extracted feature then is fed into a classifier where the regression process is performed by combining those two features using the late fusion method. finally, the classifier produces the predicted emotion dimension, which will be compared to the true value label. the difference between true value label and b t atmaja, et al. dimensional speech emotion recognition...| 94 predicted emotion dimension is the error, which is measured in three different ways. 3.1 feature extraction two sets of features from acoustic and text are used to extract emotion from speech. the following is the description of those two sets of features. 3.1.1 acoustic feature extraction a number of 31 acoustic features are used in this research. these features are, • three time-domain features: zerocrossing rate (zcr), energy, the entropy of energy. • five spectral-domain features: spectral centroid, spectral spread, spectral entropy, spectral flux, spectral roll-off. • 13 mfcc coefficients. • five fundamental frequencies (for each window). • five formants (for each window). we limit the number of windows for each utterance to 100 with 20 ms window length and 10 ms overlap. the resulting size of the acoustic feature then is (100, 31) for a single utterance. the total size of acoustic features for all utterances within the dataset is (10039, 100, 31). 3.1.2 text feature extraction text features can be obtained in many ways. one of the simple yet powerful methods is by word embedding (penningtonet al., 2014). a word embedding is a vector representation of a word. a numerical value in the form of a vector is used to make the computer to be able to process text data as it only processes numerical value. this value is the points (numeric data) in the space of dimension, in which the size of the dimension is equal to the vocabulary size. the word representations embed those points in a feature space of lower dimension (goodfellow et al., 2016). in the original space, every word is represented by a onehot vector, a value of 1 for the corresponding word, and 0 for others. this element, with a value of 1, will be converted into a point in a range of vocabulary size. to obtain a vector of each word in an utterance, that utterance in the dataset must be tokenized. tokenization is a process to divide an utterance to the number of constituent words. the following is the example of a single utterance from iemocap dataset with its tokenization and a resulted text vector for each word. text = "excuse me." tokenized_text = ["excuse", "me"] text_vector = [832, 18] to obtain the fixed length of a vector for each utterance, a set of zeros can be padded before or after the obtained vector. the size of this zeros sequence can be obtained from the longest sequence, i.e., an utterance within the dataset, which has the longest words, subtracted by the length of a vector in the current utterance. we set the longest sequence for the iemocap dataset for 554 sequences. a study to vectorize certain words has been performed by several researchers (mikolov et al., 2013; penningtonet al., 2014; mikolov et al., 2017). the vector of those words can be used to weight the word vector obtained previously. the size of the dimension of each word for pretrained word vectors is 300 (in the example above is one), shaping the size of (554, 300) text feature for each utterance, or (10039, 554, 300) for all utterances in the iemocap dataset. 95 | international journal of informatics information system and computer engineering 1 (2020) 91 102 4. dimensional emotion classifier recurrent neural network (rnn) is one of the variants of the neural network that are designed to handle sequential information. these networks introduce state variables to store past information and determine the current output based on the current input. let h is the output of hidden layer, x is the input and w is the weight of layer with bias b. then, h = φ(xwxh + bh). (1) is the output of hidden layer h with φ is nonlinear function (activation). having a recurrent hidden state (ht) whose activation at each time is dependent on that of the previous time (ht−1), the output of current hidden layer now is defined as, ht = φ(xtwxh + ht−1whh + bh). (2) the problem with that rnn is it always takes past time into consideration. a situation may be encountered when the early observation is more/less significant to predict the future. to tackle this situation and adding some enhancements, several methods have been proposed by some researchers (cho et al., 2014; hochreiter et al., 1997). in this paper, those two rnn methods are implemented as dimensional emotion classifier. 4.1 gated recurrent unit gated recurrent unit (gru) enables the gating of the hidden state in rnn. this is a mechanism that is enabled for when the hidden state should be updated and when it should be reset. it is learned and addressed some limitations of rnn e.g., whether an early observation is highly significant for predicting all future observations. if the first observation is likely of great importance, it will learn not to update the hidden state after the first observation. likewise, it will learn to skip irrelevant temporary observations. finally, it will learn to reset the latent state whenever needed (a. zhang et al., 2019). reset unit, rt, and update unit, zt, are the new additional units in gru. together with candidate unit, ĥt, it updates the gru in the following order. rt = σ(xtwxr + ht−1whr + br) (3) zt = σ(xtwxz + ht−1whz + bz) (4) h ̃ t = tanh(xtwxh +(rt ht−1) whh + bh) (5) ht = zt ht−1 + (1 − zt) h̃ t. (6) where ht now is the final update of gru rather than reset gate of the output of rnn hidden layer unit. 4.2 long short-term memory while gru using two additional units, reset and update, long short-term memory (lstm) network uses three different units to control data from current time (ht) and past time (ht−1): input, forget, and output gates. these three gates are defined as follows, it = σ(xtwxi + ht−1whi + bi), (7) ft = σ(xtwxf + ht−1whf + bf ),(8) ot = σ(xtwxo + ht−1who + bo),(9) wx ∈ d×h and wh ∈ h×h is the weight parameters with bias b ∈ 1×h the complete sequence to update hidden state is defined as follow, c̃ t = tanh(xtwxc + ht−1whc + bc) (10) ct = ft ct−1 + it c ̃ t. (11) ht = ot tanh(ct). (12) c̃ t and ct are candidate memory cell and memory cell, respectively. gru and lstm are very similar both in implementation and its result. gru is faster due to less gates and lstm, in many b t atmaja, et al. dimensional speech emotion recognition...| 96 cases, slightly better than gru due to its complexity to handle flow of data. 4.3 model architecture the implementation of the regression classifier for dimensional emotion recognition can be built by stacking some rnn layers from acoustic and text input and merge both to obtain the final dimensional emotion prediction. for each modality, acoustic, and text, we varied two dense, two gru, and two lstm layers. for rnns (gru and lstm) we also implement a bidirectional version of those networks to allow distribution of information from the past and future time (gru and lstm only roll information from the current and past time, see eq. 3–6 and eq. 10–12). those dense, bidirectional gru (bgru), and bidirectional lstm (blstm) layers from each modality is stacked together using two dense layers. fig. 2 shows one of the architectures for combining acoustic and text features to obtain three emotional dimensions. to minimize the risk of overfitting, a number of dropouts are used with value 0.4 for each acoustic and text network and 0.3 for the last dense network. rectified linear unit (relu) activation is used for both dense layers in the combined network. the final dense layer with three nodes used linear activation function to obtain the score of valence, arousal, and dominance. the whole network is trained with rmsprop (dauphin et al., 2015). optimizer with mean squared error (mse) as a loss function. beside mse, we use mean absolute error (mae) and mean absolute percentage error (mape) as evaluation metrics. the implementation of this deep learning architecture is available in public repository, https://github.com/bagustris/dimension al_ser_rnn,for research reproducibility fig. 2. architecture of deep learning system combining acoustic and text features. the number in bracket shows the size of units/nodes on the layer. 5. results 5.1 comparison of acoustic, text and combined system to begin our discussion, we presented the result for each different modality. table 1 shows the performance of dimensional speech emotion recognition from the acoustic feature. two layers of the same models are stacked and added with final a dense layer. for each model, the value of each metric is an average of five experiments (to minimize the effect of uncertainty computation due to randomness). the dense network is chosen as the baseline model. for this speech emotion recognition, the lstm model shows modest improvement from the dense baseline layer in terms of mse and mae. however, other metrics i.e., mape, shows the different result which leads gru/bgru to obtain a better result. as mse is used as a loss function, the result in mse is relevant in this context. mae metrics also show consistency with mse. the mape metric can be used for https://github.com/bagustris/ https://github.com/bagustris/dimensional_ser_rnn https://github.com/bagustris/dimensional_ser_rnn 97 | international journal of informatics information system and computer engineering 1 (2020) 91 102 comparison to other datasets as it has the same scale from 0 to 100. table 1. performance comparison of dimensional speech emotion recognition from acoustic feature (in term of mse, mae, and mape) among different models. modela mse mae mape (%) dense 0.652 0.66 24.188 gru 0.648 0.655 23.69 lstm 0.636 0.651 24.014 bgru 0.647 0.653 23.675 blstm 0.656 0.66 24.109 aeach model is a stack of two same models. for text-based emotion recognition, the result is shown in table 2. as reported by other researchers (poria et al., 2017; tripathi et al., 2018). we obtained better performance on emotion recognition for iemocap dataset by utilizing text features. in this text emotion recognition, the lstm model shows modest improvement from the baseline and other models. this result from some experiments can be considered when combining acoustic and text features for the fusion of two networks. for this text emotion recognition, all metrics show almost consistent with each other (except gru and bgru, which is 0.02 different). mae and mape show consistency in the order of the score among models. table 2. performance comparison of dimensional speech emotion recognition from text feature (in term of mse, mae, and mape) among different models modela mse mae mape (%) dense 0.493 0.559 20.436 gru 0.48 0.549 19.888 lstm 0.465 0.538 19.554 bgru 0.482 0.548 19.881 blstm 0.487 0.55 19.588 aeach model is a stack of two same models. finally, we presented the result of the fusion of acoustic and text features in table 3. clearly, a decrement of error is shown for all mse, mae, and mape metrics. for example, using the same dense layers, the error (mape) decreases from acoustic (24.188%) and text (20.436%) to combined acoustic and text system (19.97%). to obtain the more decrement of error, not only the architecture of each network modalities is important but also the strategy for combining the modalities is also important (poria et al., 2017). although we tried several different layers after concatenation of two networks (acoustic and text), we focus on selecting the combination for each modality while keeping the use of dense layer after a combination of two networks. this focus is based on some experimentation we obtained; the simple dense layers after concatenation perform better than the more sophisticated layers (gru, lstm, and attention models). table 3. performance comparison of dimensional speech emotion recognition from combination of acoustic and features (in term of mse, mae, and mape) among different models model mse mae mape (%) dense + dense 0.45 7 0.546 19.97 dense + bgru 0.44 0.533 19.585 bgru + dense 0.45 4 0.543 19.81 lstm + lstm 0.42 8 0.525 18.929 blstm + blstm 0.43 8 0.531 19.423 blstm + lstm 0.41 9 0.517 18.713 bgru + gru 0.42 9 0.527 19.139 b t atmaja, et al. dimensional speech emotion recognition...| 98 5.2 design of system architecture and its result on designing the system architecture for dimensional speech emotion recognition, we rely on initial experiments using the unimodal feature, and other researcher results (atmaja et al., 2019; tripathi et al., 2018). from tables 1 and 2, it is shown that the text feature gives better result on dimensional emotion recognition than acoustic feature. for both features, lstm performs better than any other model. using this result, we build lstmbased networks for those two modalities and combine them with dense layers. on choosing hyperparameters, we manually add more units to text networks as it gives a better result. the choice of using 50 units and 40 units of nodes on each lstm layer on each modality is also obtained from experimentation; we use larger units first and decrease this number until the smaller one without decreasing the performance (error metrics). for the dense layers, the number of 30 units for each layer is also based on the experiment. the relu and tanh activation function in those layers perform a similar result. to avoid overfitting, we use the callback strategy besides putting the dropout layer on each network branches (acoustic, text, and combination layers). two methods are used for callback (to stop the iteration of training): early stopping and model checkpointing. for early stopping, we use a number of 10 patiences to monitor validation loss. this means, if no decrement of validation loss (mse) after ten epochs, the training process will stop and uses the best weight for the evaluation/prediction. the model check pointing is a similar method to save the model (which can also be ignored if we do not want to save the model). finally, although we obtain the best prediction of emotion dimension with bltsm and lstm networks, there is a room for improvement for experimenting and designing a better model architecture. in some runs, the combination of gru performs better; however, the average result shows that a combination of lstm is the best one. the hyperparameters optimization on future research will be done on training and development set instead of manually handcrafted. for the obtained improvement, a decrement of mape from acoustic featurebased emotion recognition is achieved up to 5.5 % when using a combined feature. for mse and mae, the decrement is in a range of 0.14-0.17 and 0.090.11, respectively. from the text feature, the decrement of error is in range of 0.08-0.046, 00.02, and 0-0.84% for mse, mae, and mape, respectively. the excerpt of the result of vad score from the model obtained using blstm, and lstm networks are presented in table 4. table 4. sample of true and predicted vad score from model using blstm and lstm networks utterances vad true predicted oh, totally. yeah. [4, 3, 2.5] [3.21, 2.67, 2.60] the craziest thing just happened to me. [4, 3, 2.5] [3.35, 3.02, 2.97] this girl; she just offered me fifty thousand dollars to marry her. [3.5, 3.5, 3] [3.32, 3.3, 3.34] 99 | international journal of informatics information system and computer engineering 1 (2020) 91 102 5.3 evaluation of loss function and metrics one of the challenging problems in dimensional emotion recognition is to choose the proper metrics for evaluation. in this paper, we used standard regression metrics i.e., mse, mae, and mape. however, when running some experiments on the same condition (system architecture), when a metric decrease, an other increase the score. for example, in the second experiment, after the first, mse gets lower, but mape gets higher, and so on. table 5 shows the raw result obtained in table 1 for dense acoustic network. as shown in that table, the consistency of each metric is changing when re-running the experiment. in the second experiment, when mse score decreases, mape score increases. in the last experiment, the mse score increases, while mae and mape decrease. to evaluate metrics, we perform a simple analysis by changing the lost function from mse (default) to mae and mape. table 6 shows that by changing the loss function from mse to mae, the error result decreases slightly. table 5. results of five experiments on the same models (dense layers) from speech features experiment# mse mae mape 1 0.659 0.663 24.12 2 0.644 0.660 24.41 3 0.645 0.651 23.44 4 0.652 0.663 24.94 5 0.656 0.662 24.01 if we compare this result (our best mae) with other research which also used acoustic and linguistic information, but with different approach (karadoğan et al., 2012). our mae is better than them (their best mae is 1.28 for arousal). however, comparing the same metric across the dataset is not sufficiently comparable as the upper level bound of mae is different for each dataset. in this case, mape might be more useful than mse and mae. moreover, using another metric such as concordance coefficients correlation (ρc) as used in (tzirakis et al., 2017) is more relevant as it has the same scale 0-1 for any datasets to measure the agreement. table 6. results of blstm and lstm networks from acoustic and text features with different loss function loss function mse mae mape mse 0.43 0.523 18.87 mae 0.42 0.519 18.631 mape 0.469 0.543 18.835 b t atmaja, et al. dimensional speech emotion recognition...| 100 6. conclusion we presented our work on dimensional speech emotion recognition by combining acoustic and text features using recurrent neural networks. thirtyone acoustic features are used as input to acoustic networks, and 554word vectors are fed to text networks. the result from unimodal shows that text-based emotion recognition performs better on iemocap dataset compared to acoustic emotion recognition. the combination of acoustic and text features decreases the error of mape up to 5% from acoustic features only and near 1% from text feature only. for the combination among dnn layers, the use of blstm for acoustic network and lstm for text network with concatenated dense layers to combine those two features performs better compared to a list of given dnn layer combination. the choice of more advanced metric for loss function and evaluation in dimensional emotion recognition should be considered on the future research for consistency and benchmarking with other dimensional emotion recognition studies. references atmaja, b. t., shirai, k., & akagi, m. (2019, november). speech emotion recognition using speech feature and word embedding. in 2019 asia-pacific signal and information processing association annual summit and conference (apsipa asc), 519-523. a. zhang, z. c. lipton, m. li, and a. j. smol. (2019). dive into deep learning, http://www.d2l.ai. busso, c., bulut, m., lee, c. c., kazemzadeh, a., mower, e., kim, s., ... & narayanan, s. s. (2008). iemocap: interactive emotional dyadic motion capture database. language resources and evaluation, 42(4), 335. calvo, r. a., & mac kim, s. (2013). emotions in text: dimensional and categorical models. computational intelligence, 29(3), 527-543. chen, m., he, x., yang, j., & zhang, h. (2018). 3-d convolutional recurrent neural networks with attention model for speech emotion recognition. ieee signal processing letters, 25(10), 1440-1444. cho, k., van merriënboer, b., bahdanau, d., & bengio, y. (2014). on the properties of neural machine translation: encoder-decoder approaches. arxiv preprint arxiv:1409.1259. dauphin, y., de vries, h., & bengio, y. (2015). equilibrated adaptive learning rates for non-convex optimization. in advances in neural information processing systems 1504-1512. http://www.d2l.ai/ 101 | international journal of informatics information system and computer engineering 1 (2020) 91 102 el ayadi, m., kamel, m. s., & karray, f. (2011). survey on speech emotion recognition: features, classification schemes, and databases. pattern recognition, 44(3), 572587. giannakopoulos, t., pikrakis, a., & theodoridis, s. (2009, april). a dimensional approach to emotion recognition of speech from movies. in 2009 ieee international conference on acoustics, speech and signal processing, 65-68. goodfellow, i., bengio, y., courville, a., & bengio, y. (2016). deep learning, 1. cambridge: mit press. griol, d., molina, j. m., & callejas, z. (2019). combining speech-based and linguistic classifiers to recognize emotion in user spoken utterances. neurocomputing, 326, 132-140. hochreiter, s., & schmidhuber, j. (1997). long short-term memory. neural computation, 9(8), 1735-1780. jurafsky, d., & martin, j. h. (2014). speech and language processing, 3. karadoğan, s. g., & larsen, j. (2012, may). combining semantic and acoustic features for valence and arousal recognition in speech. in 2012 3rd international workshop on cognitive information processing (cip), 1-6. mikolov, t., sutskever, i., chen, k., corrado, g. s., & dean, j. (2013). distributed representations of words and phrases and their compositionality. in advances in neural information processing systems, 3111-3119. mikolov, t., grave, e., bojanowski, p., puhrsch, c., & joulin, a. (2017). advances in pretraining distributed word representations. arxiv preprint arxiv:1712.09405. pennington, j., socher, r., & manning, c. d. (2014, october). glove: global vectors for word representation. in proceedings of the 2014 conference on empirical methods in natural language processing (emnlp) (1532-1543). poria, s., cambria, e., hazarika, d., majumder, n., zadeh, a., & morency, l. p. (2017, july). context-dependent sentiment analysis in user-generated videos. in proceedings of the 55th annual meeting of the association for computational linguistics (volume 1: long papers) (873-883). poria, s., cambria, e., hazarika, d., mazumder, n., zadeh, a., & morency, l. p. (2017, november). multi-level multiple attentions for contextual multimodal sentiment analysis. in 2017 ieee international conference on data mining (icdm), 10331038. b t atmaja, et al. dimensional speech emotion recognition...| 102 tripathi, s., tripathi, s., & beigi, h. (2018). multi-modal emotion recognition on iemocap dataset using deep learning. arxiv preprint arxiv:1804.05788. tzirakis, p., trigeorgis, g., nicolaou, m. a., schuller, b. w., & zafeiriou, s. (2017). endto-end multimodal emotion recognition using deep neural networks. ieee journal of selected topics in signal processing, 11(8), 1301-1309. 65 | international journal of informatics information system and computer engineering 2(1) (2021) 65-76 computation in the analysis of techno-economic of the production of al2o3 (aluminum oxide) nanoparticles through precipitation method yusrianti sabrina kurniadianti*, adzra zahra ziva**, yuni kartika suryana***, risti ragadhita****, asep bayu dani nandiyanto*****, tedi kurniawan****** *,**,***,****,*****departemen kimia, universitas pendidikan indonesia, indonesia ******college community of qatar, qatar e-mail: *****nandiyanto@upi.edu a b s t r a c t s a r t i c l e i n f o this study aims to demonstrate computation in the techno-economic analysis of the production of aluminum oxide (al2o3) using the precipitation method on an industrial scale. this evaluation is based on the perspective of technical and economic evaluation. several economic evaluation parameters were analyzed to obtain potential information from the manufacture of al2o3 nanoparticles based on gross profit margin, payback period, and cumulative net present value. the results of this study identified that the manufacture of al2o3 nanoparticles using the precipitation method could be done industrially. based on the engineering perspective, al2o3 nanoparticles can be produced as much as 6.9 tons and earn an annual profit of 144,635.69 usd with a period of 20 years. to ensure that this project can be carried out, an economic evaluation is made based on estimates of ideal and non-ideal conditions, including tax increases, sales changes, raw material prices, utility prices, and labor’s salary. this study is expected to provide information for the manufacture of al2o3 nanoparticles using the precipitation method on an industrial scale. article history: ___________________ keywords: economic evaluation, alumina nanoparticles, precipitation method, al2o3 international journal of informatics, information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 2(1) (2021) 65-76 received 18 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 kurniadianti et al. computation in the analysis of techno-economic ...| 66 1. introduction nanoparticles are microscopic particles. the properties of nanoparticles with a diameter less than 100 nm differ from their bulk form (parast & morsali, 2011). nanoparticles have several uses, including catalysts (parast & morsali, 2011), anti-bacterial agents (prashant et al., 2011), anti-cancer and cell imaging (parast & morsali, 2011; prashant et al., 2011), nuclear coolant (syarif et al., 2018), and can be used in radiators or sandpaper (zhu et al., 2020) synthesis of al2o3 nanoparticles can be carried out by various methods such as sol-gel (mohammed et al., 2020; li et al., 2000); precipitation (ali et al., 2019; wang et al., 2008; hassanzadeh-tabrizi & taheri-nassaj, 2009); hydrolysis (sharifi et al., 2013); synthesis under supercritical water conditions (noguchi et al., 2008); wet chemical (lu et al., 2005; lópez-juárez et al., 2018); combustion (syarif et al., 2019; afruz & tafreshi, 2014); mechanochemical (gao et al., 2018; bodaghi et al., 2009); and microwave (hasanpoor et al., 2017). precipitation is the most efficient method for synthesizing al2o3. it is because the simplest method compared to other methods, has low raw material costs, does not pollute the environment, and has several advantages such as high purity products, high thermal stability, and the ability to control desired particle size. there are still few studies that discuss the economic evaluation of al2o3, and there is no economic evaluation based on the precipitation method. thus, the purpose of this study is to demonstrate the economic evaluation of the synthesis of al2o3 nanoparticles using the precipitation method on an industrial scale. this study contains several variations of raw materials, taxes, utilities, labor, and sales. 2. method 2.1. synthesis of al2o3 nanoparticles using the precipitation method the synthesis of al2o3 nanoparticles can be selected and improvised from the literature (ali et al., 2019; wang et al., 2008; hassanzadehtabrizi & taheri-nassaj, 2009). the materials used were alcl3.6h2o, tween80, ammonium hydroxide solution, water, and ethanol. first, alcl3.6h2o was dissolved in water to obtain a concentration of 0.4 m, and then tween80 was added to the reactor. then, ammonium hydroxide is added periodically using an injector and stirred in the reactor to produce a precipitate. after that, the precipitate was filtered, washed with water and ethanol, and the residue was taken. the residue al(oh)3 was dried in an oven at 100°c for 2 hours and then calcined using a furnace at 550°c for 5 hours to obtain γ-al2o3. then, the γ-al2o3 nanoparticles were ground in a ball mill to produce a homogeneous nano size of γ-al2o3. the process scheme of this precipitation method is shown in fig. 1. 67 | international journal of informatics information system and computer engineering 2(1) (2021) 65-76 fig. 1. al2o3 nanoparticles synthesis scheme using precipitation method 2.2 analysis method the feasibility of establishing a factory is investigated in this study from a technical and economic perspective. engineering and economic perspectives were used to evaluate the manufacture of al2o3 nanoparticles using the precipitation method by adopting references (ali et al., 2019; wang et al., 2008; hassanzadeh-tabrizi & taheri-nassaj, 2009). the method is carried out from an engineering perspective by simulating the production process on a large scale using commercially available equipment and raw materials. this process is then simulated by the mass balance that occurs during the production process. in addition, we also consider the use of raw materials and equipment to support and minimize investment costs. data evaluation is carried out from an economic perspective. it is included all equipment specifications, equipment costs, raw material prices, and utility systems. the information was gathered from online web stores such as alibaba.com. then, the data is processed using simple mathematical calculations in the microsoft excel application. the economic evaluation was calculated using several economic feasibility parameters (ragadhita et al., 2019). the kurniadianti et al. computation in the analysis of techno-economic ...| 68 economic evaluation parameters are explained briefly as follows: • gross profit margin (gpm) is the first analysis to predict a rough analysis of the profitability level of this project. gpm is calculated by subtracting the cost of selling the product and the price of raw materials. • cumulative net present value (cnpv) is a value that predicts project conditions as a function of production in years. cnpv is calculated by adding up the net present value (npv) at a particular time from the start of the project. npv is a value that expresses business expenses and income. npv itself can be calculated by multiplying the cash flow by the discount factor. • the profitability index (pi) identifies the relationship between project investment costs and the impact on business continuity or profitability. pi is calculated by dividing the cnpv by the total investment cost (tic). if the pi is less than one, then the project can be classified as unprofitable. however, if the pi is more than one, the project can be classified as a good project. • break-even point (bep) is the minimum amount of product that must be sold at a certain price to cover the total cost of production. bep is calculated by dividing fixed costs value and profits (total selling price minus total variable costs). • payback period (pbp) is a calculation to estimate the time required to return the initial capital through profits. the payback period is calculated when the cnpv/tic is at zero for the first time. the synthesis method of al2o3 nanopowder, as shown in fig. 2, is the basis for some engineering assumptions. all symbols that are shown in fig. 2 will be described in table 1. these assumptions are based on stoichiometric calculations after the project was upgraded to produce around 7 kg of al2o3 nanoparticles per cycle. the assumptions are: 1. all the chemical compositions in the reaction, such as aluminum chloride hexahydrate, surfactant tween 80, ammonium hydroxide solution, ethanol, and distilled water were used for the production of al2o3 nanoparticles. these materials are of high purity. 2. the quantity of chemicals is calculated based on the literature. 3. the scale of the chemicals is increased by up to 3000 times. 4. aluminum chloride and ammonium hydroxide are reacted in a ratio of 1: 3. 5. the al2o3 nanoparticle production process has a conversion rate of 100%. 69 | international journal of informatics information system and computer engineering 2(1) (2021) 65-76 fig. 2. process flow diagram of alumina nanoparticles. table 1. process flow diagram flow of alumina nanoparticles. no. symbol information 1 t-1 tank-1 2 p-1 pump-1 3 t-2 tank-2 4 p-2 pump-2 5 s-1 separator-1 6 o-1 oven-1 7 f-1 furnace-1 8 g-1 grinding-1 to analyze the economic perspective in this study, several assumptions are made: 1. all analysis is in usd. 1 usd = 14,500 idr. 2. the project is not supported by a bank loan. 3. all raw material prices are based on those found in the online shop (alibaba.com). aluminum chloride hexahydrate, tween 80 surfactant, ammonium hydroxide solution, and ethanol are priced at 0.4 usd/kg, 2.2 usd/kg, 1.5 usd/kg, and 0.79 usd/kg, respectively. 4. all materials used in the production process are calculated based on stoichiometric calculations. 5. the water source is free of charge because the project is located near a river. kurniadianti et al. computation in the analysis of techno-economic ...| 70 6. the total investment cost (tic) is calculated based on the lang factor (nandiyanto, 2018). 7. land purchased. therefore, the land is considered as part of the plant's initial cost and is recouped at the end of the project. 8. one cycle of making al2o3 nanoparticles takes 15 hours. 9. in a one-day process, the estimated total processing cycle is 4 cycles, assuming all tools work for continuous production based on time considerations. al2o3 nanoparticles produce as much as 29 kg per day. 10. all products are sold in full and no products are lost. 11. the al2o3 nanoparticles sell for 50 usd/kg. 12. shipping costs are borne by the buyer. 13. one year project is 240 days (the remaining days are used to clean and repair tools). 14. to simplify utility, utility units are described as units of electricity such as kwh. then, the unit of electricity is considered as a cost. assuming a utility cost of 0.099 usd/kwh. 15. total wages/labor is assumed to be fixed at 120 usd/day for 20 laborers. 16. the discount rate and annual income tax rate are 15% and 10% per year, respectively. 17. the project operation length is 20 years. an economic evaluation is carried out for a feasibility test project. this economic evaluation is done by varying the value of taxes, sales, raw materials, labor, and utilities under several conditions. tax variations are carried out at 10, 25, 50, 75, and 100%. meanwhile, variations in sales, raw materials, labor, and utilities were carried out at 80, 90, 100, 110, and 120%. 3. results and discussion 3.1. engineering perspective the process of producing al2o3 nanoparticles using the precipitation method is carried out with several instruments using industrial scales that can be obtained commercially and economically. suppose the production is carried out 960 times a year. in that case, 6.9 tons of al2o3 nanoparticles will be produced, requiring 33.00 tons of aluminum chloride hexahydrate, 60.48 tons of ammonium hydroxide, 12.21 tons of tween-80, and 200 tons of ethanol. the total price required in a year for production is 203,364.31 usd with annual sales of 348,000.00 usd, resulting in a profit of 144,635.69 usd per year. these advantages will be shown in an economic evaluation, and the value of the project will be shown over 20 years 3.2. economic evaluation 3.2.1 ideal conditions the ideal condition is shown by the analysis of the relationship between cnpv/tic and lifetime (year). cnpv/tic is on the y-axis, and lifetime (year) is on the x-axis. in the first to second year, there is a negative cnpv/tic number that is less than zero. this indicates that due to the initial cost of the production of al2o3 nanoparticles, there was a decrease in income in that year. the lowest cnpv/tic value was 71 | international journal of informatics information system and computer engineering 2(1) (2021) 65-76 0.9746 in the second year. however, in the third year, the value increased to 3.8652. this point is referred to as the payback point, and the increase shown on the graph is called the payback period (pbp). the profit to cover the initial expenses increased until it reached 33.1325 in the 21st year. moreover, the project of making al2o3 nanoparticles with the precipitation method can be considered profitable because it requires a short time to recover the investment cost, which is only 3 years. fig. 3 shows the ideal condition of the cnpv/tic graph on lifetime. fig. 3. graph of cnpv/tic on lifetime (year) under ideal conditions. 3.2.2. the effect of external conditions the external conditions of the economic evaluation influence the success of a project. the most influential external factor on the success of a project is the country's economic condition where the project is established. these factors, such as taxes levied on projects by the state to finance public spending. fig. 4 shows a graph of cnpv/tic with various tax variations over 20 years. the y-axis is cnpv/tic, and the x-axis is a lifetime (years). the pbp obtained from the tax variations of 10, 25, 50, 75, and 100% is shown in fig. 4. in the initial conditions of up to two years, the cnpv project shows negative results in various tax variations. however, in the third year, the results start to be positive, and there is a difference in the payback point of this tax variation. the payback points obtained for variations of 10, 25, 50, 75, and 100% respectively are 3.87, 3.06, 1.72, 0.38, and for 100% tax, there is no payback point because it continues to decrease every year. so, it can be concluded that the lower the tax obtained, the higher the profit obtained, and the higher the tax obtained, the lower the profit obtained. kurniadianti et al. computation in the analysis of techno-economic ...| 72 fig. 4. graph of cnpv/tic on lifetime (year) for tax variation. 3.2.3 change in sales fig. 5 shows a graph of cnpv/tic against various sales variations. the y-axis is cnpv/tic, and the x-axis is a lifetime (year). the analysis is done by increasing and decreasing the sales value by 10 and 20%, with an ideal sales value of 100%. thus, the 10 and 20% decrease is 80% and 90%, while the increase of 10 and 20% is 110% and 120%, respectively. the initial phase is in the first and second years, where sales variation is still the same because it is still in the development phase. in the third year, the effects of the project are starting to show. this can be seen from the payback points in the third year, which have positive values for variations of 80, 90, 100, 110, and 120%, namely 1.69, 2.78, 3.87, 4.95, and 6.04. a higher sales value indicates a more significant profit, and a low sales value indicates a lower profit. based on the pbp analysis, the return on investment will be obtained in 3 years when there are variations in sales of 80, 90, 100, 110, and 120%. fig. 5. graph of cnpv/tic on lifetime for sales variations 73 | international journal of informatics information system and computer engineering 2(1) (2021) 65-76 3.2.4. change in variable cost (raw material, utility, labor) changes in variable costs are caused by the prices of raw materials, utilities, and labor. fig. 6 shows the variation in raw material prices. the yaxis is cnpv/tic, and the x-axis is a lifetime (year). this analysis is done by increasing and decreasing the price of raw materials starting from 10 and 20%. when the price of raw materials is lowered, the percentages obtained are 80 and 90%. the price of raw materials is increased, the percentages obtained are 110 and 120%, and the ideal price is 100%. in the first to second years, there has been no difference in the variation in the price of this raw material. this is because the project is still in its infancy. starting to see the effect of this variation in raw material prices when entering the third year. the highest payback point is in the third year with a variation of 80%, which is 4.54, and the lowest is in the 120% variation, which is 3.19. the lower the value of the variation, the higher the profit. the higher the value of the variation will result in a lower profit. fig. 6. graph of cpnv/tic on lifetime (year) for raw material variations. fig. 7 shows the relationship of cpnv/tic to a lifetime (year) if the utility price is varied. the y-axis is cpnv/tic, and the x-axis is a lifetime (year). variations are made by adding and subtracting utility prices by 10 and 20% with a utility price value of 100%. thus, the variations obtained are 80, 90, 100, 110, and 120%. the payback points obtained in the third year are 4.1, 4, 3.9, 3.8, and 3.7. based on fig. 7, this variation in utility prices does not show a significant effect because, for 20 years, the value of cnpv/tic on the variation in utility prices is not much different. in the 20th year, the highest cnpv/tic value was shown by a variation of 80%, namely 34.5, and the lowest cnpv/tic value was indicated by a variation of 120%, which was 31.8. kurniadianti et al. computation in the analysis of techno-economic ...| 74 fig. 7. graph of cpnv/tic on lifetime (year) for utility prices variations fig. 8 shows the relationship of cpnv/tic to a lifetime (year) if variations are made on employee salaries. variations are made by adding and subtracting 10 and 20% salaries with an ideal salary value of 100%. thus, the variations obtained are 80, 90, 100, 110, and 120%. in the first and second years, there has been no difference in the variation in labor salary caused by the development period of the company. in the third year, the difference in the salary variations of these laborers began to be seen. the payback points obtained for cnpv/tic values with variations in employee salaries of 80, 90, 100, 110, and 120% are 4.2, 4.0, 3.9, 3.7, and 3.5. this utility price variation does not show a significant effect. this is indicated by the value of the variation in labor salary who are not much different. the highest cnpv/tic value in the 20th year, at 80% variation, is 35.4. meanwhile, the lowest cpnv/tic value in the 20th year, at 120% variation, is 30.9. fig. 8. graph of cpnv/tic on lifetime (year) for variations in labor salary conditions 75 | international journal of informatics information system and computer engineering 2(1) (2021) 65-76 4. conclusion based on the analysis results, the production project of al2o3 nanoparticles using the precipitation method with aluminum chloride as the main raw material shows a prospective project from an engineering and economic perspective. the precipitation method has the advantages of a simple method used, low raw material costs, and produces a high purity product. pbp analysis shows that the project can compete with the standard market because the return on investment is profitable in a short time, after about three years. from the economic evaluation analysis results, it can be concluded that this project is feasible to run. further research is needed to find ways to process waste to be reused and become environmentally friendly. 5. acknowledgements we acknowledged bangdos universitas pendidikan indonesia references afruz, f. b., & tafreshi, m. j. (2014). synthesis of γ-al2o3 nano particles by different combustion modes using ammonium carbonate. indian journal of pure and applied physics, 52(6), 378–385. ali, s., abbas, y., zuhra, z., & butler, i. s. (2019). synthesis of γ-alumina (al2o3) nanoparticles and their potential for use as an adsorbent in the removal of methylene blue dye from industrial wastewater. nanoscale advances, 1(1), 213– 218. bodaghi, m., mirhabibi, a., tahriri, m., zolfonoon, h., & karimi, m. (2009). mechanochemical assisted synthesis and powder characteristics of nanostructure ceramic of α-al2o3 at room temperature. materials science and engineering b: solid-state materials for advanced technology, 162(3), 155–161. gao, h., li, z., & zhao, p. (2018). green synthesis of nanocrystalline α-al2o3 powders by both wet-chemical and mechanochemical methods. modern physics letters b, 32(8), 1–9. hasanpoor, m., fakhr nabavi, h., & aliofkhazraei, m. (2017). microwave-assisted synthesis of alumina nanoparticles using some plants extracts. journal of nanostructures, 7(1), 40–46. hassanzadeh-tabrizi, s. a., & taheri-nassaj, e. (2009). economical synthesis of al2o3 nanopowder using a precipitation method. materials letters, 63(27), 2274–2276. li, j., pan, y., xiang, c., ge, q., & guo, j. (2006). low temperature synthesis of ultrafine α-al2o3 powder by a simple aqueous sol-gel process. ceramics international, 32(5), 587–591. lópez-juárez, r., razo-perez, n., pérez-juache, t., hernandez-cristobal, o., & reyeslópez, s. y. (2018). synthesis of α-al2o3 from aluminum cans by wet-chemical methods. results in physics, 11(november), 1075–1079. kurniadianti et al. computation in the analysis of techno-economic ...| 76 lu, h., sun, h., mao, a., yang, h., wang, h., & hu, x. (2005). preparation of platelike nano α-al2o3 using nano-aluminum seeds by wet-chemical methods. materials science and engineering a, 406(1–2), 19–23. mohammed, a. a., khodair, z. t., & khadom, a. a. (2020). preparation and investigation of the structural properties of α-al2o3 nanoparticles using the solgel method. chemical data collections, 29, 100531. nandiyanto, a. b. d. (2018). cost analysis and economic evaluation for the fabrication of activated carbon and silica particles from rice straw waste. journal of engineering science and technology, 13(6), 1523–1539. noguchi, t., matsui, k., islam, n. m., hakuta, y., & hayashi, h. (2008). rapid synthesis of γ-al2o3 nanoparticles in supercritical water by continuous hydrothermal flow reaction system. journal of supercritical fluids, 46(2), 129–136. parast, m. s. y., & morsali, a. (2011). synthesis and characterization of porous al(iii) metal-organic framework nanoparticles as a new precursor for preparation of al2o3 nanoparticles. inorganic chemistry communications, 14(5), 645–648. prashanth, p. a., raveendra, r. s., hari krishna, r., ananda, s., bhagya, n. p., nagabhushana, b. m., lingaraju, k., & raja naika, h. (2015). synthesis, characterizations, antibacterial and photoluminescence studies of solution combustion-derived α-al2o3 nanoparticles. journal of asian ceramic societies, 3(3), 345–351. ragadhita, r., nandiyanto, a. b. d., maulana, a. c., oktiani, r., sukmafitri, a., machmud, a., & surachman, e. k. a. (2019). techo-economic analysis for the production of titanium dioxide nanoparticle produced by liquid-phase synthesis method. journal of engineering science and technology, 14(3), 1639–1652. sharifi, l., beyhaghi, m., ebadzadeh, t., & ghasemi, e. (2013). microwave-assisted sol-gel synthesis of alpha alumina nanopowder and study of the rheological behavior. ceramics international, 39(2), 1227–1232. syarif, d. g., prajitno, d. h., & umar, e. (2018). characteristics of nanofluids made from solgel synthesized-al2o3 nanoparticles using citric acid and peg as organic agent and bauxite as raw material. materials science forum, 929 msf, 1–9. syarif, d. g., yamin, m., & pratiwi, y. i. (2019). self combustion synthesis of al2o3 nanoparticles from bauxite utilizing sugar as fuel for nanofluids with enhanced chf. journal of physics: conference series, 1153(1). wang, s., li, x., wang, s., li, y., & zhai, y. (2008). synthesis of γ-alumina via precipitation in ethanol. materials letters, 62(20), 3552–3554. zhu, l., liu, l., sun, c., zhang, x., zhang, l., gao, z., ye, g., & li, h. (2020). low temperature synthesis of polyhedral α-al2o3 nanoparticles through two different modes of planetary ball milling. ceramics international, 46(18), 28414– 28421. 1 | international journal of informatics information system and computer engineering 2(1) (2021) 1 14 an information sharing system for multi-professional collaboration in the community-based integrated healthcare system hideaki kanai*, akinori kumazawa** japan advanced institute of science and technology, 1 chome-1 asahidai, nomi, ishikawa 923-1211, jepang correspondence: e-mail: iida@jaist a b s t r a c t s a r t i c l e i n f o currently, japan is rapidly aging. japanese government agencies report that the percentage of elderly people whose ages are at least 65 years will increase by up to about 30 percent in 2025. as one of the measures towards this situation, the community-based integrated healthcare system will be introduced in japan. the system aims to provide elderly people living at home with appropriate health, medical, and welfare services. we focus on the burden of sharing information on the situation of the elderly at home among health, medical, welfare staffs, and neighbors. we have been developing a supporting system for sharing information on the situation of the elderly at home and conducted a field test around one year. we consider that various stakeholders involved in the community comprehensive health care system could recognize the importance of information sharing and collaboration with them through this kind of social implementation. article history: received 5 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 ___________________ keywords: index terms—ieee, ieeetran, journal, latex, paper, template. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(1) (2021) 1 14 h kanai and a kumazawa an information sharing system for...| 2 1. introduction in recent years, japan is aging. on october 1, 2018, the number of the population aged 65 and over was 35.58 million, accounting for 28.1% of the total population (aging rate) (cao, 2019). the number of elderly people living alone is remarkable. the rate of males in the elderly population was 13.3%, and that of females was 21.1% in 2015. in addition, 42.2% of men and 30.2% of women over the age of 60 want to care at home. against this background, the ministry of health, labor and welfare (mhlw) is promoting an integrated community-based healthcare system that allows older people to continue living their own lives at home and in their residential areas. the structure of the healthcare system is shown in fig. 2. the public service provided depends on the situation of the elderly. the government aims to care for the elderly at home as much as possible. the government aims to reduce welfare costs. when an older person gets sick, he goes to the hospital or admits to the hospital. when the older person feels better, he returns home. when an older person needs a daycare service, he goes to the care or admits to the care home. when the older person feels better, he returns home. in order to connect these services, organically, specialists engaged in healthcare, nursing, and welfare activities in the community should work in cooperation with each other (abe et al., 2014). the ministry of health, labor, and welfare also mentions the importance of multi-professional cooperation in-home healthcare in the community-based integrated healthcare system. in order to promote multi-professional collaboration, community care meetings are introduced in the healthcare system. the meetings aim at improving the support of the elderly and improving the social infrastructure, i.e., community improvement (john et al). namely, the care meetings play a role of sharing information of the elderly in-home care. the whole stakeholders should involve in the care meeting to share the information more detailed. however, that is not realistic because they are usually under a high workload. in this research, we aim to support multi-professional collaboration in-home care of an integrated communitybased healthcare system using an information-sharing system based on ict. we consider that the information-sharing system has advantages as follows: • grasping the situation of the elderly property. • providing the elderly with adequate care services at the proper timing. we have developed an informationsharing system for multi-professional collaboration in the community-based integrated healthcare system. in this paper, we introduce an informationsharing system for multi-professional collaboration in the community-based integrated healthcare system. in order to evaluate those issues, we conduct a field test in nomi city, ishikawa prefecture, japan. 3 | international journal of informatics information system and computer engineering 2(1) (2021) 1 14 fig. 1. rate of population aging in japan. annual report by japanese ministry of health, labour and welfare (mhlw) fy2019 2. related works currently, ict is used to support multi professional cooperation and support for caregivers. the following section introduces researches on support for multi-professional collaboration in medical care and nursing care, and support for caregivers. fig. 2. community-based integrated healthcare system in japan. annual report by japanese ministry of health, labour and welfare (mhlw) fy2019 a. multi-professional collaboration in home care amir et al. clarified the characteristics of a multidisciplinary care team, including a family doctor, a home nurse, and a neurologist, for the care of children with chronic diseases such as developmental disorders (amir et al., 2015). this study suggests that problems such as inadequate communication, such as difficulty in coordinating care between members of the care team and setting goals for treatment with family members. one of the reasons for this problem is that the activities of health care workers and caregivers are loosely coupled and weakly interdependent (pinelle et al., 2006). therefore, it is important that health care workers and caregivers can recognize each other’s activities. yamamoto et al. have developed a patient information sharing system h kanai and a kumazawa an information sharing system for...| 4 among multi-professional healthcare professionals for home healthcare collaboration (yamamoto et al., 2018). this system has a patient information sharing function, healthcare worker information sharing function, and text chat. these functions are useful for sharing information among multiprofessional medical staff belonging to different medical institutions. this study shows medical staff belonging to core hospital elderly people have shown interest in information other than records registered by healthcare professionals belonging to different organizations, and the possibility of smooth medical cooperation through the notification function. this system is not intended for family or patient use. therefore, it is considered difficult to obtain requests and consultations from family members and to obtain home information from the family’s point of view. in this study, the information sharing system using ict is used not only by medical and nursing care workers but also by the elderly family members. our system aims that communication between medical professionals and family members could be activated. b. support for family caregivers there are some researches that use ict to support caregivers. duncan et al. developed a monitoring system for elder care [7]. the system installs sensors and cameras at the entrance to the elderly’s residence. the system takes a photograph of the person at the entrance when the doorbell rings, or open and sent it as an image message to the caregiver’s mobile phone. the system enables caregivers to reduce the cognitive load for confirming the elderly going out and visiting people from remote places. as a result, they improve the care and quality of life for the elderly. as psychological support for family caregivers, yamashita et al. clarified the stress nature and needs of family caregivers with depressed patients and their demands for social relationships with others (yamashita et al., 2013). based on these findings, they developed a web application for nursing records to support family caregivers of depressed patients (yamashita et al., 2017). family caregivers can use the web application to record daily activities and moods of depressed patients, record the activities of family caregivers, and facilitate family caregiver reflection analysis. this application clarified that the communication between patients and family caregivers could be activated. in this study, information that is usually difficult to know, such as the activity and physical condition of the elderly at home, is shared not only with health care professionals but also with especially distant families. 3. initiatives for homemedical cooperation of the community-based integrated healthcare system in the target field in this paper, we introduce an information-sharing system for multiprofessional collaboration in the community based integrated healthcare system. we conduct a field test of the system in nomi city, ishikawa prefecture, japan. the following is the introduction of nomi city and its home medical cooperation. 5 | international journal of informatics information system and computer engineering 2(1) (2021) 1 14 a. overview of nomi city, ishikawa prefecture as of 2015, nomi city, ishikawa prefecture, had a population of 48,881, and the population aged 65 and over is 11,983 (stat.go.jp, 2020). according to the japan medical association, the aging rate in nomi city, ishikawa prefecture in 2015, was 24.5% (national average 26.3%). the number of doctors per 100,000 populations is 147.30 (national average 230.56), and the number of nurses and nursing staff is 29.56 (national average 18.17) and 109.33 (national average 93.02), respectively. the total number of nursing homes is about 15% higher than the national average. in nomi, ishikawa, there are more caregivers than medical staff. b. multi-professional cooperation among home-basedmedical care in nomi city in this research, we collaborate with a community care conference on “medical care and nursing care cooperation” organized by a promotion organization of multiprofessional cooperation in nomi: “memory care network nomi.” we investigate problems in-home medical care cooperation and develop and introduce an informationsharing system. the activities of this “medical and nursing care cooperation” are classified into two categories, “multi-occupation coordination meeting” and “home care utilizing ict”. “multi-professional coordination meeting” is divided into “regional meeting” and “on-site meeting.” at the regional meetings, various stakeholders, such as doctors, caregivers, and elderly support center staff participate. at the regional meeting, activity reports in the three districts of nomi city, reports on the contents of onsite meetings, and discussions on the operation and functions of an information-sharing system using ict will be held. in the on-site meeting, family caregivers participate in addition to the stakeholders in each of the field districts. they confirm their opinions and reports on what they want to share. the on-site meeting is held once a month. in “home care using ict,” elderly people in target field districts are monitored using the information-sharing system introduced. the contents monitored by each professional are input to the system and shared among professionals. 4. information-sharing system: mcnbooksystem a. system overview we have developed an informationsharing system “mcnbooksystem” to support multi-professional collaboration in the community-based integrated healthcare system. we conduct a field test in nomi city, ishikawa prefecture, japan. fig. 3 shows the configuration of this system. mcnbooksystem aims to facilitate the sharing of information among medical professionals of different affiliated organizations and family members. medical and nursing care workers, including family caregivers, input information they have noticed from the state and conversation of the elderly into the system and share the information with other professionals using devices such as smartphones, tablets, and personal computers via the internet. the items to be shared are as follows. each item is entered as an arbitrary level. for example, a status of temperature is represented as “extreme fever and need h kanai and a kumazawa an information sharing system for...| 6 to contact with medical professional,” “high fever,” “a little fever rather than normal,” “nomal.” (fig. 2). • physical condition: temperature (4 levels), blood pressure (4 levels), dullness (4 levels) and excretion status (4 levels) • intake status: drug (3 levels), meal (4 levels) and water (3 levels) • activity: going out (4 levels), exercise, independence of cooking (4 levels), independence of excretion (4 levels) and exercise (3 levels) • living situation: sleep (4 levels), communication (4 levels), and garbage disposal (4 levels) • features on individuals: items according to the characteristics of the target. ex.,levels of back-pain • involved member: family members, contact list and stakeholders fig. 3. system structure b. functions for the information-sharing system the system supports multi-professional collaboration in home care by the following functions. • function of sharing elderly situation • function of sharing messages • function of altering changes in elderly status in the function of sharing elderly situation, users can input and access the shared items mentioned above. in order to make it easier for users to enter and confirm the status of the elderly, the information is displayed and shared in four stages: “emergency”, “caution,” “attention” and “normal” (fig. 5). users also can trace the history of these statuses (fig. 6). in the function of sharing messages, users can communicate with messages among them like sns (fig. 7). users get to know the status of the elderly except for shared items, and advice and instructions from others. users also can attach image files on each message in order to share handwritten care record documents, diseased area and activity (fig. 8). in the function of alerting changes in elderly status, the system can alert the pre-registered persons when there are changes in the situation of the elderly. the criteria of the alert are based 7 | international journal of informatics information system and computer engineering 2(1) (2021) 1 14 on some rules of an individual. for example, some statuses became “emergency,” and the situation continues after fifth observation in succession. fig. 4. screen image: inputting shared system fig. 5. screen image: each status of the elder displayed as four stages: “emergency”, “caution,” “attention” and “normal” h kanai and a kumazawa an information sharing system for...| 8 fig. 6. screen image: history of statuses of the elderly fig. 7. screen image: sharing messages like sns 5. a filed test of the system we conducted a field test of the system in three districts of nomi city. the number of subjects are six elderly persons. a. subjects the subjects of the test are elderly people in 70s and 80s. the start of operation of the system differs each subject due to each subject’s circumstances. table 1 shows the subject’s information on the kick-off date of operation, the household situation, the level of care needed, the degree of independence of the elderly with dementia, and stakeholders of inhome care. subject a and b live in the first district, subject c and d live in the second district, and subject e and f live in the third district. b. system usage logs in the field test 9 | international journal of informatics information system and computer engineering 2(1) (2021) 1 14 we analyzed system usage logs for three districts where this system has been installed. there are two subjects in each district. the responsible persons are a doctor “doctor a” who is in charge of the first district, a doctor “doctor b” who is in charge of the second district, and a doctor “doctor d” and a care manager ”care manager xx” who are in charge of the third district. table 1, table 2 and table 4 show usage logs whose stakeholders for subjects in each district. they are the totaled results of usage logs from the start of using the system until september 30th, 2017. for example, table ii shows that there are one doctor “doctor a,” three family members “family-1, family-2 and family-3,” one care center, one care manager, one care house, and one home helper for subject a, and these usages logs. according to table ii and table iv, there are some users who had never used the system since the number of logins is 0. on the other hand, tthere are most of the users who tended to confirm the status of subjects and messages frequently. we found that households living together tended to share information due to “emergency” or “caution ”, and households living alone tended to share messages. 6. interview survey we interviewed with system users of the filed tests. through the interview, we investigated the following things: • situation of multi-professional cooperation before the introduction of the system • changes in care and work after the introduction of the system. • changes in the elderly after the introduction of the system the following results are based on stakeholders of “subject a.” “doctor a” is a family doctor of the subject. table 1. usage logs whose stakeholders for subjects in each district subject (age) kick-off date household situation level of care needed (5 levels) degree of independence of the elderly with dementia (8 levels) stakeholders of in-home care a:80s may 31th,2017 alone 3 none family,doctor,care manager, center,care house b:80s nov 1st,2016 child, wife of child 1 6 family,doctor,care manager, center,care house (2 places) c:80s may 8th, 2017 wife, cild 3 7 family,doctor,care manager, center,care house (2 places), neighbor d:70s dec 26th, 2016 wife 1 3 family, care manager, center,care house, neighbor e:80s dec 20th, 2016 alone 1 none family,doctor,care manager, center,care house f:80s dec 12th, 2016 alone 1 none family,doctor,care manager, center,care house h kanai and a kumazawa an information sharing system for...| 10 fig. 8. attached image files on each message: handwritten care record documents, diseased area and activity a. situation of multi-professional cooperation before the introduction of the system “doctora”commented,“theonlywaytoco ntactother professionals was to contact them by phone or in-person instead of email, and the information of the elderly alone was available only during consultation hours. ” and the doctor commented, “as a result, i did not know the medical information from other medical institutions, and caregivers or care managers may decide the care policy and service without consulting with me.” “care manager a” commented, ”care managers have a central role and must contact with other stakeholders frequently. however, the only time i can meet the family doctor is the elderly’s consultation hours, and the way to know the situation of the elderly is only a monthly report from the nursing care establishment. hence, i cannot discuss it with other professionals elaborately.” moreover, the care manager commented, “i have many elderly persons whom i am in charge of, and have to visit, especially those living alone twice a week. i have to spend much time and effort to understand what elderly people with dementia say. i usually contact some issues of elderly persons with a care center once a month. i end up to deal with them even if i report them.” “family-1” commented, “before using the system, we visit the elderly without staying a once a week or ten days since we live a distance from the elderly. the actual situation was usually different from what i had heard about the elderly and the house from the elderly. we frequently contact the family doctor and the care manager difficulty due to time 11 | international journal of informatics information system and computer engineering 2(1) (2021) 1 14 issues, and we are vaguely anxious about this situation.” according to these comments, we find that it was difficult to contact each other before the introduction of the system. and we consider that information on the elderly could only be obtained incompletely among the whole stakeholders. the stakeholders can perform appropriate interventions in the elderly difficultly. the families consulted with the professionals and did not know the appropriate response. hence, we consider that inappropriate care policies and services could be set possibly. b. changes in care and work after the introduction of the system the care house commented, “we have only information about what the elderly act in the care house twice a week. using the system, we can get to know the meals and medicines at home and understand that what we need to pay attention to.” the home helper commented, “i know the situation at home, and thoughts of the elderly’s family more clearly. in a case that i would like to know the situation of the elderly at home, which other professionals have, i had no choice but asking the care manager about it. the situation information becomes available to me. at my staff meeting, i can inform other staff members of the information other professionals have, which could not be shared.” the care manager commented, ”it is effortless to realize what to should care for and tell the family since we can access the current situation and information of the elderly in the system. as a care manager, the number of visits to the subject decreased. i used to visit a family doctor, care house, and the elderly’s household; however, now, i only need to go to the care house or the elderly’s household once a week”. the family commented, ”my younger brother’s wife is also checking on the system, and she can come to the house at a time when she is worried. she can know the proper timing to visit.” in the logs of ”subject a” in table ii, the whole stakeholders frequently use the system. according to the number of checking the information, they emphasize to sharing information among them. “family-1” commented, ”there is a note of caring for the elderly in a care house. i have to visit the care house in order to contact the note. using the system, i can check the note in the system. i realize that the staffs care for the elderly carefully. i should input the current situation of the elderly firmly. i would like to input the state of the elderly of the house and the contents of the conversation with the elderly as much as possible. i am not worried about the elderly because a family doctor and other professionals can care for the elderly and give some advice depending on the current situation of the elderly”. from these comments, we consider that the active monitoring by multi-professionals was a kind of “trigger” that encouraged family members sharing information. c. changes in the elderly after the introduction of the system we find that “subject a” changed after the introduction of the system.” for example, “family a” repeatedly said to “subject a” that many people keep observing and caring for you and observing various. as a result, it seems that subject a gradually began to be aware of this situation. “family a” commented, “there are some sensors to recognize whether “subject a” lives a regular life based on his motion or not. in a case that the sensors do not recognize it, the sensors alert to the home helper. h kanai and a kumazawa an information sharing system for...| 12 before the introduction of the system, the alter often occurs. since “subject a” changes to live a regular life” after the introduction of the system, the number of alerts decreased. “subject a” commented: “i feel that family and other people watch over. so, i began to live a regular life.” this change in the subject is one of the good effects on the elderly person. we consider that the system fosters multi-professional collaboration, and the quality of care improves. 7. discussion in the following, we discuss the effects of using the mcnbooksystem on multiprofessional collaboration in-home care and future directions based on these considerations. a. effects on multi-professional collaboration in-home care the analysis using the logs and interview surveys reveals that the effects of cooperation among the whole stakeholders including family members using the system as follows: it is possible to grasp the state of the elderly at home, which had been difficult to confirm, to find out and share problems unique to the elderly. • it is possible to obtain detailed information of the elderly through the function of information sharing of this system and reduce the effort of care: i.e., the number of visits among the professionals and the elderly’s homes. the result shows that professionals could care for more elderly people. • it is useful for reducing the burden on care managers who play a central role in cooperation among other professionals, including family members, and improving efficiency by utilizing information from them. • the family members were encouraged to input and share the state of the elderly at home because they understand that they could grasp the exact situation of the elderly and obtain the appropriate advice from the professionals through the system. • the input of the information by the whole stakeholders led to the reduction of their burden and the improvement of care. thus, we consider that the informationsharing by the whole professionals not only improves care for the elderly, but also has the effect of reducing work load and increasing efficiency, and promoting the sharing of information with the elderly’s families. b. difference in the usage status of this system for each household compared to the first district in table 2 and the third one in table 4, the use of this system seems to be less in the second one in table 3. “care manager of subject a” commented that “multi-professional collaboration using such an informationsharing system would not be successful if only medical professionals used.” we consider that one of the reasons is that there is little information from family members of the elderly. comparing table 2 and table 3, the use of the system by family members of “subject c” is less than the use by ones of “subject a” even though family members of subject c started operation of this system at the same time as ones of subject a. we consider that there is a relation ship between the usage by the family 13 | international journal of informatics information system and computer engineering 2(1) (2021) 1 14 members and one by professionals. namely, as the use by the family increase, the information-sharing by the whole stakeholders would activate. therefore, we consider that it is important to regard family members as members of multiprofessionals and to actively share information with each other in multi professional collaboration. doctor b inputs the situation more frequently than other doctors but has less input and confirmation of messages. doctor a stated: ”the number of inputs depends on the purpose of the usage. if a doctor who has a strong relationship with medical information inputs a message, a care manager who wants the information will inevitably input and confirm messages. we consider that the sharing of medical information by doctors is one of the factors of active information sharing. c. effects and effects on the elderly we consider that the system fosters multi-professional collaboration using information sharing among the whole stakeholders. we find that the system also had a positive effect on the elderly who were watched over. for example, subject a said, “i feel that family and other people watch over. so, i began to live a regular life.” we consider that the system provides the elderly with a kind of awareness of health. d. issues of the system we consider that specifying the degree of importance enables users to grasp essential messages at a glance, and contributes to more precise information sharing. on the other hand, it is not easy for stakeholders, especially family members, to set appropriate importance for each message. it is necessary to carefully consider whether they will avoid the input operation itself due to the complexity of the work and the increase in the amount of work. in a case that high importance set in a message, the stakeholders might have a kind of sense of duty toward dealing with it urgently. this sense increases such a burden to the stake holders that they might avoid using message-sharing. we consider the balance between the improvement of the function and the increase in the burden. in the future, we will pay careful attention to the impact on the field use in order to improve 8. conclusion in this paper, we introduce an information-sharing system to support multi-professional collaboration in the community-based integrated healthcare system we conducted a filed test of the system in three districts of nomi city, ishikawa prefecture, japan. according to an analysis of the system usage logs and an interview survey, it is found that the information-sharing system was effective in reducing the burden on professionals such as care managers and improving efficiency. it is also found that active information sharing by professionals promotes activities of confirming and inputting messages for elderly family members. also, the elderly were aware that the stakeholders watch over them carefully. the awareness had the effect of making their daily life regular. from these results, we consider that the use of the system makes it possible to share the information which was difficultly shared among the whole stake holders, and shows the possibility of promoting the health of the elderly themselves. the system. h kanai and a kumazawa an information sharing system for...| 14 acknowledgment this research was supported by the japan science and technology agency(jst), research institute of science and technology for society (ristex)(jpmjrx16g3). we would like to thank the memory care network nomi for their kind cooperation in the questionnaire survey and interviews in this study. references “annual report on the ageing society [summary] fy2019,” https://www8.cao.go.jp/kourei/english/annualreport/2019/ pdf/2019.pdf (access feb 10th, 2020) abe, y., & morita, t. (2014). a measure to quantify the quality of communication and cooperation among medical, nursing, and welfare services in a region. palliative care research, 9(1), 114-120. https://www.jstage.jst.go.jp/article/jspm/9/1/9_114/ _pdf, (access feb 10th, 2020)) amir, o., grosz, b. j., gajos, k. z., swenson, s. m., & sanders, l. m. (2015, april). from care plans to care coordination: opportunities for computer support of teamwork in complex healthcare. in proceedings of the 33rd annual acm conference on human factors in computing systems, 1419-1428. care meeting. http://www.mhlw.go.jp/seisakunitsuite/bunya/ hukushi_kaigo/kaigo_koureisha/chiiki-houkatsu/dl/link3-0-02. pdf john duncan l. jean camp william r. hazelwood. the portal monitor: a privacyenhanced event-driven system for elder care persuasive’09article no.362009 pinelle, d., & gutwin, c. (2006). loose coupling and healthcare organizations: deployment strategies for groupware. computer supported cooperative work (cscw), 15(5-6), 537-572. yamamoto, r., yoshino, t., nishibata, m., nakai, k., yanagimoto, m., & irie, m. (2018). evaluation of a cooperative patient information system for home care providers, 59(5), 1351–1361. yamashita, n., kuzuoka, h., hirata, k., kudo, t. (2013). understanding the conflicting demands of family caregivers caring for depressed family members chi’13 2637-2646. yamashita, n., kuzuoka, h., hirata, k., kudo, t., aramaki, e., & hattori, k. (2017, may). changing moods: how manual tracking by family caregivers improves caring and family communication. in proceedings of the 2017 chi conference on human factors in computing systems, 158-169. http://www.stat.go.jp/english/index.html (access feb 10th, 2020 http://www.stat.go.jp/english/index.html%20(access 1 | international journal of informatics information system and computer engineering 2(1) (2021) 83124 practical computation in the techno-economic analysis of the production of magnesium oxide nanoparticles using sol-gel method jessica veronica*, lidia intan febriani*, citra nurhashiva*, risti ragadhita*, and asep bayu dani nandiyanto*, tedi kurniawan** *departemen pendidikan kimia, fakultas pendidikan matematika dan ilmu pengetahuan alam, universitas pendidikan indonesia, jl. dr. setiabudi no. 229, bandung, 40154, indonesia **community college of qatar, qatar a b s t r a c t s a r t i c l e i n f o the purpose of this study was to determine the feasibility of a project for the manufacture of magnesium oxide nanoparticles using the sol-gel method by evaluating both technically and economically. evaluation from the technical side is determined by stoichiometric calculations and evaluation of the initial factory design, while the evaluation from the economic side is determined by several parameters to determine the benefits of the project to be established (gross profit margin, internal rate return, break-even point, payback period, and cumulative net present values). some of these economic evaluation parameters were analyzed to inform the production potential of magnesium oxide nanoparticles, such as determining the level of profitability of a project (gross profit margin), predicting the length of time required for an investment to return the initial capital expenditure (payback period), predicting the condition of a production project in the form of a production function in years (cumulative net present value), etc. the results of the technical analysis show that this project can produce 1,425 kg of magnesium oxide nanoparticles per day and the total cost of the equipment purchased is 45,243 usd. payback period analysis shows that the investment will be profitable after more than three years. to ensure article history: received 5 nov 2021 revised 20 nov 2021 accepted 25 nov 2021 available online 26 dec 2021 __________________ keywords: economic evaluation, magnesium oxide nanoparticles, sol-gel method international journal of informatics information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 2(1) (2021) 1-12 jessica et al. practical computation in the techno-economic...| 2 project feasibility, the project is estimated from ideal to worst-case conditions in production, including salary, sales, raw materials, utilities, as well as external conditions such as taxes. 1. introduction magnesium oxide is an important functional metal oxide that has been widely used in various fields, such as catalysts, refractory materials, paints, and superconductors (dobrucka, r. 2018). magnesium oxide nanoparticles are highly ionic metal oxide nanoparticles with a very high surface area and unusual crystal morphology (dobrucka, r. 2018). magnesium oxide nanoparticles have been widely used because of their unique properties, such as large band gap, thermodynamic stability, low dielectric constant, and low refractive index (prasanth, r., et al. 2019). several methods can be used in the synthesis of magnesium oxide nanoparticles, including combustion (balakrishnan, g., et al. 2020), synthesis of plant extracts (essien, e. r., et al. 2020), sonochemical synthesis (yunita, f. e., et al. 2020), solid-state synthesis (zhang, h., et al. 2019), and sol-gel synthesis (taghavi fardood, s., et al. 2018). of these several methods, the most appropriate method for conducting economic evaluation analysis is the sol-gel synthesis method that has been carried out by fardood, et al (taghavi fardood, s., et al. 2018). the sol-gel method is one of the most preferred methods for synthesizing magnesium oxide nanoparticles because of its simple process, high product yield, and low reaction temperature. in addition, sol-gel is an inexpensive method to get magnesium oxide nanoparticles with narrow size distribution and larger surface area which is very important to solve the problems of low reactivity and catalytic ability (mguni, l. l., et al. 2013). figure 1. shows a diagram of the manufacture of magnesium oxide nanoparticles using the sol-gel method with arabic gum as the raw material. 3 | international journal of informatics information system and computer engineering 2(1) (2021) 83124 fig. 1. schematic of the sol-gel method of manufacturing magnesium oxide nanoparticles many studies have described various methods of synthesizing magnesium oxide nanoparticles, but there are no studies that have studied the economic evaluation of large-scale synthesis of magnesium oxide nanoparticles. therefore, this study aimed to analyze the project economy of manufacturing magnesium oxide nanoparticles using the sol-gel method on an industrial scale. evaluation is carried out from two sides, such as the technical side and the economic side. on the technical side, it can be determined by stoichiometric calculations and evaluation of the initial factory design, while the evaluation from the economic side is determined by several parameters to determine the benefits of the project to be established (gross profit margin, internal rate return, break-even point, payback period, and cumulative net present value) under various conditions (nandiyanto, a. b. d., 2018). 2. method in this study, we chose the research on the manufacture of magnesium oxide nanoparticles conducted by fardood, et al (taghavi fardood, s., et al. 2018) as the main reference. in the economic evaluation, an analysis of the prices of equipment, utilities, and raw materials available for the manufacture of magnesium oxide nanoparticles was carried out on the alibaba online shopping website. then, the data is calculated using microsoft excel with several parameters, such as gross profit margin, internal rate return, break-even point, payback period, and cumulative net present value of various cost variables. the calculations were carried out based on the literature (nandiyanto, a. b. d., 2018), (ragadhita, r., et al. 2019), (nassar, m. y., et al. 2017), (garrett, d. e., 2012). to get the results of this study, calculations are carried out using several formulas such as: ● gross profit margin (gpm) is the first analysis to determine the level of profitability of a project. this analysis is estimated by reducing the cost of selling the product with the cost of raw materials. gpm = 𝛴𝑡𝑟=1 𝑡𝑟 (𝑆 . 𝜂 − 𝑅𝑀)𝑃𝐶 . 𝑄 . 𝑡 (1) s is the total sales, rm is the total raw materials, pc is the production capacity, q is the capacity of raw materials that are included and used in the process (kg/hour), and t is the production time. ● internal rate return is a presentation that describes the average interest profit per year from all expenses and income of the same amount. if the internal rate return is greater than the prevailing real interest (current jessica et al. practical computation in the techno-economic...| 4 bank loan interest), then the factory is considered profitable, but if the internal rate return is less than the prevailing real interest (current bank loan interest), then the factory is considered a loss. npv = 𝛴𝑡𝑟=1 𝑡𝑟 𝐶𝑜 (1+𝑖)𝑡𝑟 – co (2) co is the total investment cost, ct is the net cash inflows during the period, tr is the project time (in years), and i is the discount rate that can be obtained in alternative investments. ● break-even point (bep) is the minimum amount of product value that must be sold at a certain price to cover the total cost of production. break-even point can be calculated by dividing fixed costs by profit. ● payback period (pbp) is a calculation to predict the length of time required for an investment to return the initial capital expenditure. in short, the payback period is calculated when the cumulative net present value reaches zero. ● cumulative net present value (cnpv) is the total value of net present value (npv) from the beginning of the factory construction until the factory ends operation. npv = 𝛴𝑡𝑟=1 𝑡𝑟 ( 𝑅𝑡 (1+𝑖)𝑡𝑟 ) (3) 3. results and discussion 3.1. procedure in this study, several assumptions were used based on the illustration of the process of making magnesium oxide nanoparticles shown in figure 2. this assumption shows that by increasing the project through stoichiometric calculations, approximately 1,425 kg of magnesium oxide nanoparticles is produced in one cycle. the assumptions are: (1) all raw materials are upgraded to 500,000 times of the lab-scale in the literature. (2) the ingredients are of high purity. (3) magnesium nitrate hexahydrate and arabic gel solution are reacted and produce magnesium oxide with a purity of 98%. (4) the loss during the process of moving, drying, and collecting the product is 2%. there are several assumptions used to ensure economic analysis. this assumption is needed to analyze and predict several possibilities that occur during the project. the assumptions are: (1) all analyzes use usd (1 usd = 14,383 rupiah) (bank indonesia, 2021); (2) based on commercially available prices, the prices of arabic gum and magnesium nitrate hexahydrate are 2.44 usd/kg and 0.26 usd/kg, respectively. all materials are estimated based on stoichiometric calculations; (3) when project land has been purchased, land costs are added at the beginning of the factory construction year and recovered at the end of the project; (4) total investment cost (tic) is calculated based on lang factor (garrett, d. e., 2012); (5) total investment cost is prepared in at least two steps. the first step is 40% in the first year and the second step is the remainder (during project development); (6) depreciation is estimated using direct calculation (garrett, d. e., 2012); (7) one cycle of the manufacturing process for magnesium oxide nanoparticles takes 16 hours; (8) the cost of postage shall be borne by the buyer; (9) magnesium oxide 5 | international journal of informatics information system and computer engineering 2(1) (2021) 83124 nanoparticles are sold at 2 usd/pack (1 kg); (10) one year project is 300 days (and the remaining days are used to clean and organize the process); (11) to simplify utility, utility units can be described and converted as electrical units, such as kwh (nandiyanto, a. b. d., 2018). then, the unit of electricity is converted into charge. the unit of electricity (kwh) is multiplied by the cost of electricity. the assumed utility cost is 0.078 usd/kwh; (12) the total salary/labor is assumed to be at a fixed value of 68.36 usd/day; (13) the discount rate is 15% per year; (14) income tax is 10% annually; (15) the length of operation of the project is 10 years. economic evaluation is carried out to test the feasibility of the project. this economic evaluation is carried out by varying the value of raw materials, utilities, sales, salary, and taxes under several conditions. variations in raw materials, utilities, sales, and salary were carried out at 50, 75, 100, 125, 150, 175, and 200%. tax variations are carried out at 10, 25, 50, 75, and 100%. fig. 2. illustration of the flow diagram for the manufacture of magnesium oxide nanoparticles. table 1. table of a process flow diagram for the manufacture of magnesium oxide nanoparticles. no symbol information 1 r-1 reaktor-1 2 r-2 reaktor-2 3 pu-1 pump-1 4 pu-2 pump-2 5 fi-1 filtrasi-1 jessica et al. practical computation in the techno-economic...| 6 6 fu-2 furnace-1 7 g-1 grinding-1 figure 2. shows the process of making magnesium oxide nanoparticles using the sol-gel method using arabic gum as raw material based on the literature of fardood, et al (taghavi fardood, s., et al. 2018). all the symbols in figure 2 are informed in table 1. first, arabic gum was dissolved with distilled water in the reactor for 120 minutes at 75oc to reach a clear arabic gel solution and then transferred to the next reactor. after that, magnesium nitrate hexahydrate was added to the arabic gel solution in the reactor, and stirring was continued for 14 hours to obtain a brown resin. the liquid was filtered, then continued into the furnace and the final product was calcined at 550˚c for 4 hours. after that, the magnesium oxide sample was pulverized with a special mechanical powder smoothing tool to obtain the nanoparticle size (taghavi fardood, s., et al. 2018). one cycle produces 1,425 kg of magnesium oxide nanoparticles. in one month, the project can produce 35,625 kg and in one year the project can produce 427,500 kg of magnesium oxide nanoparticles. from an engineering perspective, the total cost for purchasing raw materials for one year is 263,674 usd, while the total sales in one year are 256,500,000 usd. the profit for one year was 256,236,325 usd. the price for the equipment cost analysis is 45,243 usd. total investment cost must be less than 191,830 usd. the life of the project is 10 years, producing magnesium oxide nanoparticles with cumulative net present value/total investment cost reaching 3,789.411%, in the tenth year and in the third year the payback period has been successfully achieved. 3.2. economic evaluation 3.2.1. ideal condition figure 3. shows a graph of the relationship between cumulative net present value/total investment cost with respect to time. the y-axis is the cumulative net present value/total investment cost and the x-axis is the lifetime (year). the curve shows a negative cumulative net present value/total investment cost (%), which is a value below 0 in the first year to the third year, which indicates a decrease in revenue in that year due to the initial capital cost for the production of magnesium oxide nanoparticles. in the third year the graph shows an increase in income, this condition is the payback period (pbp). profits can cover the initial capital that has been spent and continue to increase thereafter until the tenth year. in table 2. the cumulative net present value/total investment cost is negative from the first year to the second year. then the value of cumulative net present value/total investment cost began to return to positive in the third year. thus, the production of magnesium oxide nanoparticles can be considered a profitable project because it requires a short time to recover the investment costs. 7 | international journal of informatics information system and computer engineering 2(1) (2021) 83124 fig. 3. the ideal condition for cumulative net present value/total investment cost to a lifetime (year). table 2. annual cumulative net present value under ideal conditions. cnpv/tic year 0 0 -0,4093519278 1 -0,8452045511 2 733,6408078950 3 1372,3242969786 4 1927,7012440078 5 2410,6377196854 6 2830,5824811441 7 3195,7518389344 8 3789,4109083098 9 3789,4109083098 10 3.2.2. the effect of external conditions the success of a project can result from external factors. one factor is the taxes levied on projects by the state to finance various public expenditures. figure 4. shows a graph of cumulative net present value with various tax variations. the yaxis is cumulative net present value/total investment cost (%) and the x-axis is a lifetime (year). figure 4. shows that the conditions from the beginning of the year to the second year show the same results because the cumulative net present value is under tax variations and there is project development. in addition, in that year there was no income and there was a reduction in accordance with the graph of ideal conditions. profits continue to increase after reaching the payback period (pbp) until the tenth year. cumulative net present value/total investment cost in the tenth year for each variation of 50, 75, 100, 125, 150, 175 and 200% is 41.68; 41.05; 39.99; 38.94, and 37.89%. so, the maximum tax for earning a break-even point (the point at which there is both profit and loss in the project) is 75%. tax changes of up to more than 75% lead to failure in the project. fig. 4. cumulative net present value curve of tax variations. 3.2.3. change in sales figure 5. shows a graph of cumulative net present value with various sales variations. the y-axis is the cumulative net present value/total investment cost and the x-axis is the lifetime (year). the results of the payback period are shown in figure 5. the conditions from the beginning of the year to the second year of the cumulative net present value project in various variations are the same. this happened because of the project development. the effect of sales on jessica et al. practical computation in the techno-economic...| 8 cumulative net present value can be obtained after the project is made for 2 years from the initial conditions. the greater the value of the sale, the more the profits obtained from the project are carried out. however, if there are conditions that cause product sales to decline, the project's profits will decrease from ideal conditions. based on the payback period (pbp) analysis, the payback period for sales variations of 50, 75, 100, 125, 150, 175, and 200% can be achieved in the third year. profits continue to increase after reaching the payback period until the third year. the profit margin generated for each year increases with increasing sales from ideal conditions. cumulative net present value/total investment cost in the tenth year for each variation of 50, 75, 100, 125, 150, 175, and 200% is 1890.71; 2840.06; 3789.41; 4738,77; 5688.11; 6637.47; and 7586.82%. so, the minimum sale to get the break-even point (the point at which the project's profit or loss) is 50%. sales of magnesium oxide nanoparticles will be more profitable if sales are increased by more than 50% because the graph shows a positive cumulative net present value/total investment cost, this means the project is feasible (nandatamadini, f., et al. 2019). fig. 5. cumulative net present value curve of sales variation 3.2.4. change in variable cost (cost of raw material, salary, utility) there are several internal factors such as the condition of the cost of raw materials, salary, and utilities that can affect the success of a project. figure 6. shows a graph of cumulative net present value with various variable costs of raw materials. the y-axis is the cumulative net present value/total investment cost and the x-axis is the lifetime (year). the analysis is done by lowering and increasing the cost of raw materials by 25, 50, 75, and 100%. the ideal cost of raw materials is 100% when the cost of raw materials is reduced by 25 and 50%, the cost of raw materials becomes 75 and 50%, respectively. when the cost of raw materials is increased by 25, 50, 75, and 100%, the cost of raw materials will be 125, 150, 175, and 200%. the payback period is obtained from the variable cost of raw materials. the results of the payback period are shown in figure 6. the conditions from the beginning of the year to the second year of the cumulative net present value project in various variable costs of raw materials are the same. it is because of the project development. the effect of the cost of raw materials on the cumulative net present value can be obtained after the project is made for 2 years from the initial conditions. the lower the cost of raw materials, the higher the profit of the project. however, if there are circumstances that cause the cost of raw materials to increase, the project profit will decrease from the ideal situation. based on payback period analysis, profits continue to increase after reaching the payback period (pbp) until the tenth year. however, the profit margin obtained every year is getting smaller 9 | international journal of informatics information system and computer engineering 2(1) (2021) 83124 with the increase in the cost of raw materials from ideal conditions. on the other hand, the annual profit margin increases with a decrease in the cost of raw materials from ideal conditions. cumulative net present value/total investment cost in the tenth year for each variation of 50, 75, 100, 125, 150, 175, and 200% is 3789.60; 3788.55; 3787.50; 3786.45; 3785.40; 3784.35; and 3783.30%. from the variable cost of raw materials, the project can still run and make a profit. fig. 6. cumulative net present value of the variable cost of raw materials. figure 7. shows a graph of cumulative net present value with various salary variations. the y-axis is the cumulative net present value/total investment cost and the x-axis is the lifetime (year). the analysis is done by increasing and decreasing salary by 25, 50, 75, and 100%. the ideal salary is 100%. when the salary is reduced by 25 and 50%, the salary will be 75 and 50% respectively. when the salary is increased by 25, 50, 75, and 100%, the salary will be 125, 150, 175, and 200%. the payback period is obtained from the results of salary variations. the results of the payback period are shown in figure 7. the conditions from the beginning of the year to the second year of the cumulative net present value project from various salary variations are the same. this happened because of the project development. the effect of salary on cumulative net present value can be obtained after the project is made for 2 years from the initial conditions. there is no significant change from the salary variation curve to the cumulative net present value graph. the payback period for each salary variation is still achieved in the third year. however, the cumulative net present value/total investment cost differs in the tenth year for each variation. the difference in values for each variation of 50, 75, 100, 125, 150, 175, and 200% is 3790.37; 3789,89; 3789.41; 3788.93; 3788.45; 3787.98; 3787.50%. from the salary variations, the project can still run and make a profit. fig. 7. cumulative net present value curve of salary variations figure 8. shows the cumulative net present value graph with various utility variations. the y-axis is the cumulative net present value/total investment cost and the x-axis is the lifetime (year). the analysis is done by increasing and decreasing the utility by 25, 50, 75, and 100%. the ideal utility is 100%, when the utility is reduced by 25 and 50%, the utility becomes 75 and 50%, respectively. when the utility is increased by 25, 50, 75, and 100% then the utility becomes 125, 150, 175, and 200%. the payback period is obtained from the results of utility jessica et al. practical computation in the techno-economic...| 10 variations. the results of the payback period are shown in figure 8. the conditions from the beginning of the year to the second year of the cumulative net present value of various utility variations are the same. this is because of the project development. the effect of utility on cumulative net present value can be obtained after the project is made for 2 years from the initial conditions. there is no significant change from the utility variation to the cumulative net present value graph. however, the cumulative net present value/total investment cost differs in the tenth year in each variation. the difference in values for each variation of 50, 75, 100, 125, 150, 175, and 200% is 3789.57; 3789.49; 3789.41; 3789.33; 3789.25; 3789.17; and 3789.08%. the payback period for each utility variation is still achieved in the third year. from the utility variations, the project can still run and make a profit. fig. 8. cumulative net present value curve of utility variations 4. conclusion based on the analysis that has been carried out, the project to manufacture magnesium oxide nanoparticles from an engineering point of view shows that the scale of the project can be scaled up using currently available tools and has a relatively low cost. payback period analysis shows that the investment is profitable after more than three years. this is because the use of arabic gum as a raw material in the synthesis of magnesium oxide nanoparticles by the sol-gel method is cheap and environmentally friendly. from this economic evaluation analysis, it can be concluded that this project is feasible to run. acknowledgments we acknowledged bangdos, universitas pendidikan indonesia. references balakrishnan, g., velavan, r., batoo, k. m., & raslan, e. h. (2020). microstructure, optical and photocatalytic properties of mgo nanoparticles. results in physics, 16, pp. 103013. bank indonesia, “foreign exchange rates”. [online]. available: https://www.bi.go.id/id/statistik/informasi-kurs/transaksibi/default.aspx, 2021, retrieved august 12, 2021. 11 | international journal of informatics information system and computer engineering 2(1) (2021) 83124 dobrucka, r. (2018). synthesis of mgo nanoparticles using artemisia abrotanum herba extract and their antioxidant and photocatalytic properties. iranian journal of science and technology, transactions a: science, 42(2), pp. 547-555. essien, e. r., atasie, v. n., okeafor, a. o., & nwude, d. o. (2020). biogenic synthesis of magnesium oxide nanoparticles using manihot esculenta (crantz) leaf extract. international nano letters, 10(1), pp. 43-48. garrett, d. e. (2012). potash: deposits, processing, properties and uses. springer science & business media. mguni, l. l., mukenga, m., jalama, k., & meijboom, r. (2013). effect of calcination temperature and mgo crystallite size on mgo/tio2 catalyst system for soybean oil transesterification. catalysis communications, 34, pp. 52-57. nandatamadini, f., karina, s., nandiyanto, a. b. d., & ragadhita, r. (2019). feasibility study based on economic perspective of cobalt nanoparticle synthesis with chemical reduction method. cakra kimia (indonesian e-journal of applied chemistry), 7(1), pp. 61-68. nandiyanto, a. b. d. (2018). cost analysis and economic evaluation for the fabrication of activated carbon and silica particles from rice straw waste. journal of engineering science and technology, 13(6), pp. 1523-1539. nassar, m. y., mohamed, t. y., ahmed, i. s., & samir, i. (2017). mgo nanostructure via a sol-gel combustion synthesis method using different fuels: an efficient nano-adsorbent for the removal of some anionic textile dyes. journal of molecular liquids, 225, pp. 730-740. prasanth, r., kumar, s. d., jayalakshmi, a., singaravelu, g., govindaraju, k., & kumar, v. g. (2019). green synthesis of magnesium oxide nanoparticles and their antibacterial activity. indian journal of geo marine sciences, 48(08), pp. 12101215. ragadhita, r. i. s. t. i., nandiyanto, a. b. d., maulana, a. c., oktiani, r. o. s. i., sukmafitri, a. j. e. n. g., machmud, a. m. i. r., & surachman, e. (2019). “techno-economic analysis for the production of titanium dioxide nanoparticle produced by liquid-phase synthesis method,” journal of engineering science and technology, 14(3), pp. 1639-1652. taghavi fardood, s., ramazani, a., & woo joo, s. (2018). eco-friendly synthesis of magnesium oxide nanoparticles using arabic gum. journal of applied chemical research, 12(1), pp. 8-15. yunita, f. e., natasha, n. c., sulistiyono, e., rhamdani, a. r., hadinata, a., & yustanti, e. (2020, june). time and amplitude effect on nano magnesium jessica et al. practical computation in the techno-economic...| 12 oxide synthesis from bittern using sonochemical process. in iop conference series: materials science and engineering, 858(1), p. 012045. zhang, h., hu, j., xie, j., wang, s., & cao, y. (2019). a solid-state chemical method for synthesizing mgo nanoparticles with superior adsorption properties. rsc advances, 9(4), pp. 2011-2017. 103 | international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 swot analysis of lending platform from blockchain technology perspectives wisnu uriawan insa lyon, liris, umr5205, france department of informatics, uin sunan gunung djati bandung, indonesia correspondence: e-mail: wisnu.uriawan@insa-lyon.fr, wisnu_u@uinsgd.ac.id a b s t r a c t s a r t i c l e i n f o blockchain technology become phenomenal issue in the world, emerging with bitcoin and iot. it had been implemented in many areas in human activity. advantages of blockchain technology is distributed ledger where resources distributed to all member in network. loans or credit as part of human activity in their life. when we need cash for a major expense, it might be tempting to borrow from a payday lender or max out a credit card or similar, but you have other options that will not harm your credit or put you in a cycle of debt, even if your credit record is not all that great. collateral loans could be a way to borrow the money as you need. one of which is the lending platform. blockchain technology has been implemented in many lending platforms, but there are still any weaknesses that can be refined and optimized. tool of analysis is swot, describes of four analysis, as follow: strength, weakness, opportunity and threats. this paper purpose to analyzed lending platform measured by look for the weaknesses variable and how to optimize that system performance that can improve for helping people in lending process. in addition, the result of this analysis can use for enhancement recommendation the system based on the weaknesses are found and opportunity for make a lending platform robust. article history: received 18 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 available online 09 sep 2018 ___________________ keywords: blockchain, distributed ledger, lending platform, swot, enhancement, robust. international journal of informatics, information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 w uriawan. – swot analysis of lending platform from blockchain technology perspectives |104 1. introduction blockchain arise in good time, it introduced with bitcoin emerging in financial crisis in 2008 (g. verdian, p. tasca, c. paterson, and g. mondelli,, 2018), at the same time people need some money for survive in life. traditional lender burdened people with high interest and more long-time payback period. another way, individual can propose lending to traditional financial services, like a bank or financial institutions (i. gonzález-carrasco, j. l. jiménezmárquez, j. l. lópez-cuadrado, and b. ruiz-mezcua,2019), but it’s very difficult for accepted because they should give some collateral. on the other hand, people do not have collateral for proposed some credit to financial institutions, more big proposed loans, will more collateral also (w. presthus and n. o. o’malley,2017). financial institutions make some conditions for grading lending, it depend on the risk. high risk will high conditions and more collateral required. blockchain has completed by technology modern for support in lending platform environment with the simple way and faster in processing (i. anagnostopoulos,2018). ledger will support for recorded all transactions on network with unique privilege it is work as individual or organization as well as (j. li, d. greenwood, and m. kassem,2019). smart contracts are supported to manage transactions with particular process with high code security and robustness with solidity adaptive language (w. viriyasitavat and d. hoonsopon,, 2019), peer-to-peer (p2p) lending is possible for person transaction with another person online with simple way, in several country has successful apply that system as like in china and united kingdom, its method provides for small medium enterprises (smes) and individual with population increasing about 110% in year 20152016 and still developing for lending industry market with high risk (y. zhang, h. li, m. hai, j. li, and a. li, 2017), trustworthiness will help the transaction more faster because will reduced many following conditions like transaction credit record, collateral, many person rejected when applied lending or credit with small collateral ( c. r. macaulay 2015), ( t. yu and w. shen 2019). according to the transparency market research, lending market cap will growth in 2024 about $897.85 then before in 2015 about $26.16 billion (b. bilbao and v. argentaria, 2018) is significant and interesting for lending platform development in the future. all the kind of services will provide with blockchain technology for financial institutions non-bank. peerto-peer networking a set group computer or network that can resource sharing by itself autonomous, without central authority need support by consensus. asymmetric cryptography used for ensure the user can accesses the system and avoidance from malicious access or unauthority users, with this cryptography model only priviledge and authentic user can access the system (j. chen and s. micali, 2019). cryptographic hashing merkle tree data structure is used to record all transactions from large datasets with unique authority so that data integrity is well maintained (l. wang, x. shen, j. li, j. shao, and y. yang, 2019). swot analysis identify several factors of 105 | international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 organizational or system for adapted and analyze present conditions for competitive advantages. strength will identify positive value of the current process, weakness it shown vulnerability or error of the current system according to process after organizational evaluated. opportunity use for identifying in the future beside competitive advantages. threats is how the system can avoidance or minimized the risk today and in the future (s. s. humaidan, 2015), swot analysis will help the organization for manage better all the situation and conditions occurs also for increasing competitive advantages. this paper describes how to analyze the lending platform by looking for particularly the weaknesses and opportunity, that can improve the lending platform performance. in addition, weaknesses show the system leak or vulnerable so that the organization can fix this weakness become lending platform strength and robust. 2. methodology blockchain is a coherent data structure as a digital sign and becomes an identity, then shared and distributed as a database that contains a wellstructured log of transaction and chronological information. the database is called a ledger which contains many transactions that have occurred, user logging and data maintenance. transactions in the general ledger will be collected into a block, which is recorded with a time stamp and cryptographically linked to the genesis block that forms and composes a sequence of events or other blocks as a unit. the structure itself describes the data structure, is a form of digital consensus and is also used in literature, algorithms or application domains that are built on the design block structure (m. andoni et al, 2019), (j. mattila et al, 2016), (g. wood, 2014). data transmission, sharing of resources, and computing are all part of computer networks. cryptocurrency as a distributed network resource to others with a specific address. the challenge is that the system needs to ensure that the expenses that are distributed do not double. a traditional transaction such as bank and non-bank, which acts as a mediator between a third party and a trusted data storage medium, will become a valid ledger block and keep the data up to date. authority activated if several parties need to write in the ledger at the same time with concurrency control and consolidation. the current centralized management may not accept this, because it is high cost and requires the trust of network users to a third party to operate the system (m. andoni et al, 2019), (j. mattila et al, 2016). 2.1 ledger a copy of the distributed ledger will be the same held by all users on the network. consensus will add new data every time there is a change to the ledger if all users agree and the data is valid. any attempt to change the data by one user then the other users will be informed of the change, it creates immutability (s. pearson et al, 2019). the genesis block is the place where transactions are stored, and all blocks w uriawan. – swot analysis of lending platform from blockchain technology perspectives |106 will be cryptographically linked to each other which records the data. the nodes in the graph represent transactions and their edges will show the direction of confirmation between transactions (b. shala, u. trick, a. lehmann, b. ghita, and s. shiaeles, 2019). 2.2 fintech the financial technologies (fintech) and the iot (i. anagnostopoulos, 2018), (b. shala, u. trick, a. lehmann, b. ghita, and s. shiaeles, 2019), ( j. l. ferrer-gomila, m. francisca hinarejos, and a. p. isern-deyà, 2019). although both domains require the basic common features of a decentralized trusted transaction ledger, substantial differences can be found in usage principles, transaction volume and rates, tight security requirements or transaction fees. at fintech, the main challenge is ensuring highly secure and reliable financial payments, with a low volume of transaction failures and some tolerance for transaction delays. on the other hand, in iot, using multiple devices and a greater volume of transactions is expected, with micro and nano payments required for iot assets and data changes. transaction costs are a relevant issue here, as well as transaction delays required for near real-time operations (m. pustišek and a. kos, 2018), shown by fig. 1. fig. 1. lending process 2.3 ethereum ethereum is a blockchain network with an open system. ethereum and bitcoin are tools that make it possible to bring the economic system into the software, complete with an account management system and a native exchange unit system for funneling through accounts (l. catania, s. grassi, and f. ravazzolo, 2019). games like monopoly or something identic. each player calls a native unit of coins for this exchange, token, or cryptocurrency, but it is no different from tokens on any other system: they are a form of money (or scripts) that can only be used in that system (c. dannen, 2017). ethereum is a global transactionbased machine: it starts with block genesis and gradually executes transactions to manipulate it into final blocks. the last block is canonical ethereum. genesis blocks available in it such as account balance information, track records, reliable, real-world data and information; otherwise, anything that could be manipulated by the computer would be accepted. transaction validation between blocks is essential to avoid invalid fake blocks or valid blocks undergoing changes. invalid block changes will reduce the account balance without the same 107 | international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 renewal and opposite increases elsewhere. valid block transitions generated by transactions (u. team, 2018). decentralized protocols and applications, distributed storage, distributed shared marketplace and other concepts have the potential to enhance the computing industry and the support provided for peer-to-peer protocols for lending platforms. writing code in ethereum, will get a lot of blockchain functionality like a programming language completely useful for generating smart contracts functionality and more secure by arbitrary encoding and users can modify according to the objective (b. v. buterin, 2009). 2.4 smart contract the assigned and delegated smart contracts will execute the correct input. distributed ledger is a distributed database of transaction history that is approved by the majority of participants in its network through a predetermined consensus mechanism. all participants in the network have a copy of the same ledger. any changes in the ledger will be reflected in all on the final copy. assets that are recorded in the ledger can be financial, legal, physical, electronic, or many other properties. depending on the network rules, the ledger can be updated by some or all the participants. major security and consensus issues are generally resolved through cryptographic mechanisms (a. raschendorfer et al, 2019). smart contracts are computer programs that enforce rules without requiring a third party. in the bitcoin blockchain, basic versions of smart contracts are implemented through a writing system which can facilitate use cases such as multi-user accounts, multi wallets, digital signatures and other services (m. westerkamp, f. victor, a. küpper, and a. kupper, 2019). the security risks associated with blockchain platforms and technology are as follows: 1) it is difficult to implement regulations to avoid money laundering activities due to the high level of anonymity using the protocol. 2) there are several cryptocurrencies that have been indicated to be used for money laundering purposes and are a challenge ahead to enforce financial penalties. 3) the ineffective tax regulation scheme may be a side effect of the anonymity factor. 4) many countries have introduced tax regulations for cryptocurrencies, in an effort to deal with the global cryptocurrency phenomenon and avoid risks for users who are exposed to cryptocurrency price volatility. 5) in the future, there will be problems of inflation and risks of monetary and financial stability, including the potential loss of control over the amount of currency to be circulated, the level of risk will increase from the cryptocurrency market and increased credit in the economy with the use of cryptocurrencies that are difficult to control. 6) the emergence of the use of cryptocurrency will be accompanied by a series of new problems that arise with regard to laws and regulations, so that it can affect public confidence in conventional currencies. cryptocurrency is like a representation of value transactions, can be traded and function generally as money (d. unal, w uriawan. – swot analysis of lending platform from blockchain technology perspectives |108 m. hammoudeh, and m. s. kiraz, 2019). society of worldwide interbank financial telecommunication (swift) system uses predefined code to pass on transaction details through the swift network. each transaction is described by a series of swift code. the code consists of several key identifier components, such as institution code, country code, location code and branch code to indicate the sender and receiver (t. qiu, r. zhang, and y. gao, 2019). swot analysis it contains about relation identification of strengths, weakness, opportunities and threats from subject research. strength contra with weakness, opportunity versus threats. strength will identify about resources available and running well, weakness for measure how high the risks. opportunity analyze represent the marketplaces as the subject, while threats is for prevent or minimize the risk ( t. qiu, r. zhang, and y. gao, 2015), ( j. r. glass, g. h. kruse, and s. a. miller, 2015), ( h. thamrin and e. w. pamungkas, 2017), ( a. aich and s. k. ghosh, 2016). identifying the lists of positive and negative contribution objectives with the subject evaluation. swot analysis is required for the selection of right method and application, is very simple but the result is better, and the recommendation can use for next process. 3. results and discussion blockchain has become famous and is present for all daily activities in the fields of economy, business, education, health, and other non-bank financial institutions. exponential growth has almost changed the world and provides more options for efficiency and low costs. the lending activity is an interesting area because many people have experiences at least once in their life. in practice of lending system, someone will become a lender and another party will be a borrower. there are many terms and conditions for this loan activity, including: collateral, ability to payback, trust, and the purpose of the loan. blockchain technology will bridge these activities to reduce transaction times and complicated mechanisms. in the traditional lending process, it will take at least 1 week to a month, but borrowers cannot wait longer because of urgent needs. 3.1. analysis lending platform smart contracts in blockchain technology have been used to replace a faster and more efficient lending process, one example of changing from centralization to a process of decentralization, trust, efficiency and accountability. one of the spanish multinational banking groups banco bilbao vizcaya argentaria (bbva) invested around € 75 million in corporate loans and has taken advantage of the distributed ledger and became the first international bank to implement blockchain technology .swot analysis for lending platform variable are blockchain technology, no mediator/collateral, competitive rates, reduced risk, improved 109 | international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 efficiency, market cap and competitive advantages. swot analysis framework summarized (b.bilbao and v. argentaria, 2018), (t. qiu, r. zhang, and y. gao, 2019) in the following table 1. table 1. swot analysis framework variable strengths weakness opportunitie s threats blockchain technology √ no mediator/ collateral √ competitive rates √ trustiness √ reduced risk √ improve efficiency √ market cap √ competitive advantages √ the lending platform will help individuals who are looking for a place to apply for loans or credit directly to lenders. this process will reduce dependence on financial institutions such as banks and third parties. blockchain technology adopts the traditional lending process by reducing waste and making decisions quickly. illustration "if someone needs a loan or credit, they just visit one of the loan platforms and try to propose an amount of credit need, then the system will response with status of accepted or rejected" in just a few minutes. blockchain technology is helping for better and more profitable transactions such as: safer, simpler operations, generating potential passive income, and people are starting to focus on blockchain development. some loan platforms offer convenience to use, but they require a lot of terms and conditions. however, it is still a burden for users, if the lending platform still requires collateral, users will have difficulty fulfilling it so that users will still be rejected. finally, they tried to return to traditional loans. investment in a lending platform is very interesting as long as it is completed with new technology that allow it to meet with the user experience and needs, many lending platforms only focus on profit without thinking about the user needs. lending platforms offer the lending products but do not focus on user needs analysis so that will affect the success of a lending platform. 3.2. analysis lending platform current lending platforms able to implement of blockchain technology through third parties, bitcoin and ethereum. referring to the swot analysis shown in table 2. w uriawan. – swot analysis of lending platform from blockchain technology perspectives |110 bitcoin has a limitation that the bitcoin network will be a threat to future lending platforms and is quite difficult to develop. ethereum is still open to development, therefore it is easier to develop and will grow faster without worrying about restrictions. third parties, it is easier to start installing because they are developed by third parties, but it will be high cost and also requires third party maintenance. mix (bitcoin and ethereum), the opportunity is open to take the lower to upper segment classes, but the system is more complex to handle both transactions. 3.3. trustworthiness the most important thing about the lending platform is the level of trust, if the community has given trust, the lending platform will grow faster and have a good influence so that it will be recommended to other users. the best practices of each user will be to transfer his experience, refer to table 3 below. ease of use describe how far lending platform have acceptance from users without complaint and satisfy in use. track record acceptance, statistical record acceptance and succeed for lending process. member complaint, variable strengths weakness opportunitie s threats bitcoin √ ethereum √ third-party √ mix (bitcoin and ethereum) √ bitcoin √ ethereum √ third-party √ mix (bitcoin and ethereum) √ variable strengths weakness opportunitie s threats ease of use √ track record acceptance √ member complaint √ time needed √ market capital √ table 2. swot technology platform framework table 3. swot of trustness 111 | international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 reduction or minimize of member complaint will indicate the good process. time needed, how long lending process will need processing time, time that's shown can cut down complex process. market capital, how big market penetration of lending platform, larger coverage area will bring trustiness to users. 3.4. improved efficiency efficiency is a keyword of the lending platform performance parameters will show in the table 4 below. ratio of use, lending platform usability will describe the overall system performance. comparison of lenders and borrowers, providing statistical data between lenders and borrowers so that users will learn the comparison of the two and see future trends. asset vs market, the lending platform must be able to present the asset report and market capitalization to the public, the user will study and analyze the results, for example, the asset must be greater than the market capitalization to guarantee the user's investment. optimized technology, the loan platform must always update the system periodically to keep all transactions up and running efficiently. accumulation of errors, to minimize errors, is the obligation of the loan platform to ensure the system runs well with zero errors. 3.5. market capital each lending platform will competitive with many lending platforms in the world, they should prepare variable in table 5 below. variable strengths weakness opportunitie s threats ratio of use √ comparative of lender and borrower √ asset vs market √ technology optimized √ error accumulation √ variable strengths weakness opportunitie s threats performance √ speed √ technology driven √ security √ table 4. swot of improve efficiency table 5. swot of competitive advantages w uriawan. – swot analysis of lending platform from blockchain technology perspectives |112 performance, the lending platform must have a good performance to be the winner or the best choice for the user. speed is an indicator to measure how fast each transaction can be handled and measure the success of the transaction process. driven by technology, the latest technological innovations will help improve performance and in line with the system architecture, the more recent technology uses will affect the lending platform's adaptability to the latest technology. security, a classic problem but very important to apply to the lending platform, a good system must have a high level of security, because it will be accessed by many users. 3.6. competitive advantages growth of lending platform shown by market capital, variables of market capital refers to table 6 below. coverage area, how big is the coverage area of the lending platform that will be supported, the wider area will give opportunity to win the market with the risk of being more complex in managing the system. kind of country, the type of user will be influenced by the type of the state and shows the behavior of the user, the system must be able to handle all the behavior of users who access. market segmentation will show the user segment environment, middle to upper or lower class will be used as evaluation for market segmentation, corporate or personal trends will also measure the level of communication of the lending platform. lending variants, lending platform products that are available in various variants and segmentations can open a wider market. 3.7. comparative of lending platform comparative of lending platform will describe how performance their system was built and how the system can handle many transactions with many products are offer refers to table 7 below. variable strengths weakness opportunitie s threats coverage area √ kind of country √ market segment √ variant of lending √ table 6. swot of market capital 113 | international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 4. conclusion the blockchain technology of the lending platform will affect the overall system performance, swot analysis variables show 6 strengths, 5 weaknesses, 6 opportunities and 5 threats. all variables describe how to manage the lending platform at the best possible performance. these recommendations will help in the future to develop a better lending platform and system maintenance considering the strengths and opportunities, weaknesses and threats variables. however, for the best lending platform, focus on weaknesses and threats to make the system more reliable. 5. acknowledgements authors wishing to acknowledge liris umr5205 laboratory at insa lyon france and mora scholarship from indonesian government that variable strength weakness opportunity threats blockchain technology bitcoin √ ethereum √ third-party √ mix (bitcoin and ethereum) √ reduced risk easy of use √ track record acceptance √ member complaint √ time needed √ market capital √ improve efficiency ratio of use comparative of √ lender and borrower √ asset vs market √ technology optimized √ error handle √ technology optimized coverage area √ kind of country √ market segment √ variant of lending √ error accumulation performance √ speed √ technology driven √ security √ table 7. swot of competitive advantages w uriawan. – swot analysis of lending platform from blockchain technology perspectives |114 supports and funds this research publication. references aich, a., & ghosh, s. k. (2016). application of swot analysis for the selection of technology for processing and disposal of msw. procedia environmental sciences, 35, 209-228. anagnostopoulos, i. (2018). fintech and regtech: impact on regulators and banks. journal of economics and business, 100, 7-25. andoni, m., robu, v., flynn, d., abram, s., geach, d., jenkins, d. & peacock, a. (2019). blockchain technology in the energy sector: a systematic review of challenges and opportunities. renewable and sustainable energy reviews, 100, 143-174. b. bilbao and v. argentaria, “top blockchain projects related to loans : salt lending crypto-collateral lending,” 2018. b. v. buterin, (2009). offsetmapping, 1–36. brad, s., & brad, e. (2015). enhancing swot analysis with triz-based tools to integrate systematic innovation in early task design. procedia engineering, 131, 616-625. c. dannen. (2017) introducing ethereum and solidity. catania, l., grassi, s., & ravazzolo, f. (2019). forecasting cryptocurrencies under model and parameter instability. international journal of forecasting, 35(2), 485501. chen, j., & micali, s. (2019). algorand: a secure and efficient distributed ledger. theoretical computer science, 777, 155-183. ferrer-gomila, j. l., hinarejos, m. f., & isern-deyà, a. p. (2019). a fair contract signing protocol with blockchain support. electronic commerce research and applications, 36, 100869. glass, j. r., kruse, g. h., & miller, s. a. (2015). socioeconomic considerations of the commercial weathervane scallop fishery off alaska using swot analysis. ocean & coastal management, 105, 154-165. gonzález-carrasco, i., jiménez-márquez, j. l., lópez-cuadrado, j. l., & ruiz-mezcua, b. (2019). automatic detection of relationships between banking operations using machine learning. information sciences, 485, 319-346. humaidan, s. s. (2015). an exploratory study to determine the suitability of applying information systems using (swot) analysis, at administrative affairs department in power plant maintenance centre. procedia computer science, 65, 787-796. 115 | international journal of informatics information system and computer engineering 4 (1) (2019) 103-116 li, j., greenwood, d., & kassem, m. (2019). blockchain in the built environment and construction industry: a systematic review, conceptual models and practical use cases. automation in construction, 102, 288-307. macaulay, c. r. (2015). capitalism's renaissance? the potential of repositioning the financial ‘meta-economy’. futures, 68, 5-18. mattila, j., naucler, c., stahl, r., tikkanen, m., & bâdenlid, a. (2016). industrial blockchain platforms: an exercise in use case development in the energy industry (no. 43). etla working papers. pearson, s., may, d., leontidis, g., swainson, m., brewer, s., bidaut, l., ... & zisman, a. (2019). are distributed ledger technologies the panacea for food traceability?. global food security, 20, 145-149. presthus, w., & o’malley, n. o. (2017). motivations and barriers for end-user adoption of bitcoin as digital currency. procedia computer science, 121, 89-97. pustišek, m., & kos, a. (2018). approaches to front-end iot application development for the ethereum blockchain. procedia computer science, 129, 410-419. qiu, t., zhang, r., & gao, y. (2019). ripple vs. swift: transforming cross border remittance using blockchain technology. procedia computer science, 147, 428434. raschendorfer, a., mörzinger, b., steinberger, e., pelzmann, p., oswald, r., stadler, m., & bleicher, f. (2019). on iota as a potential enabler for an m2m economy in manufacturing. procedia cirp, 79, 379-384. shala, b., trick, u., lehmann, a., ghita, b., & shiaeles, s. (2019). novel trust consensus protocol and blockchain-based trust evaluation system for m2m application services. internet of things, 7, 100058. thamrin, h., & pamungkas, e. w. (2017). a rule based swot analysis application: a case study for indonesian higher education institution. procedia computer science, 116, 144-150. u. team, “trustworthy peer financing with ultiledger blockchain protocol and ecosystem,” 2018. unal, d., hammoudeh, m., & kiraz, m. s. (2020). policy specification and verification for blockchain and smart contracts in 5g networks. ict express, 6(1), 43-47. verdian, g., tasca, p., paterson, c., & mondelli, g. (2018). quant overledger whitepaper. viriyasitavat, w., & hoonsopon, d. (2019). blockchain characteristics and consensus in modern business processes. journal of industrial information integration, 13, 3239. wang, l., shen, x., li, j., shao, j., & yang, y. (2019). cryptographic primitives in blockchains. journal of network and computer applications, 127, 43-58. w uriawan. – swot analysis of lending platform from blockchain technology perspectives |116 westerkamp, m., victor, f., & küpper, a. (2020). tracing manufacturing processes using blockchain-based token compositions. digital communications and networks, 6(2), 167-176. wood, g. (2014). ethereum: a secure decentralised generalised transaction ledger. ethereum project yellow paper, 151(2014), 1-32. yu, t., & shen, w. (2019). funds sharing regulation in the context of the sharing economy: understanding the logic of china's p2p lending regulation. computer law & security review, 35(1), 42-58. zhang, y., li, h., hai, m., li, j., & li, a. (2017). determinants of loan funded successful in online p2p lending. procedia computer science, 122, 896-901. 117 | international journal of informatics information system and computer engineering 1(1) (2020) 117-128 international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 1(1) (2020) 117 128 computation application: techno-economic analysis on the production of magnesium oxide nanoparticles by precipitation method lidia intan febriani*, citra nurhashiva**, jessica veronica***, risti ragadhita****, asep bayu dani nandiyanto*****, tedi kurniawan****** *,**,***,****,*****departemen pendidikan kimia, universitas pendidikan indonesia, indonesia ******college community of qatar, qatar e-mail: *****nandiyanto@upi.edu a b s t r a c t s a r t i c l e i n fo this study aims to analyze the feasibility of a project for the production of magnesium oxide nanoparticles using precipitation methods on a large scale. the feasibility analysis of this project was determined using an evaluation from an economic and engineering perspective. evaluation from an engineering perspective is determined by the evaluation of the initial factory design and stoichiometric calculations. meanwhile, the evaluation from an economic perspective is determined by several parameters, such as payback period, gross profit margin, cumulative net present value, etc. the analysis results show that the production of magnesium oxide nanoparticles using the precipitation method can be carried out on an industrial scale. in this project, 11,250 kg of magnesium oxide nanoparticles were obtained per day, and the total profit earned was 1,881,184,752.91 usd in 10 years. payback period analysis shows that the investment will be profitable after more than three years. to ensure project feasibility, projects are estimated from ideal to worst-case conditions in production, including salary, sales, raw materials, utilities, and external conditions such as taxes. article history: received 17 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 keywords: economic evaluation, magnesium oxide nanoparticles, precipitation method mailto:nandiyanto@upi.edu febriani et al. computation application: techno-economic analysis on the production…| 118 i. introduction magnesium oxide (mgo) is one of the most useful ceramic materials because it has a high melting temperature (around 2800oc) (tai et al., 2007). magnesium oxide is commonly used for catalysis and remediation of toxic wastes or as an additive in paints (ding et al., 2001). magnesium oxide also has good additive properties, which can be applied to high fuel oils (agrawal et al., 2015). magnesium oxide is needed in industrial, environmental, and health products. magnesium oxide is used industrially as a catalyst in many applications includes organic carbonate synthesis catalysts (kantam et al., 2007). another application of magnesium oxide is in steel manufacture because it is highly corrosion-resistant (z zhang et al., 2015). magnesium oxide nanoparticles can be produced through several synthesis methods. synthetic methods that can be used in the synthesis of magnesium oxide nanoparticles include combustion (balakrishan et al., 2020), synthesis of plant extracts (essien et al., 2020), sonochemical synthesis (yunita et al., 2020), solid-state synthesis (zhang et al., 2019), sol-gel synthesis (taghavi et al., 2018), and precipitation (alvionita and astuti, 2017). of the several methods, the precipitation method is one of the most preferred methods for the synthesis of magnesium oxide nanoparticles because it allows control over the particle size so that the time required is relatively short and can be carried out at low temperatures (alvionita and astuti, 2017). therefore, the precipitation method was chosen as the method to be analyzed through economic evaluation in producing magnesium oxide on an industrial scale. fig. 1 shows a diagram of the manufacture of magnesium oxide using the precipitation method. fig. 1. schematic of the process of making magnesium oxide nanoparticles using the precipitation method. previously, there have been many studies describing the process of synthesizing magnesium oxide nanoparticles using the precipitation method. however, there is no study that examines the economic evaluation of the synthesis of magnesium oxide nanoparticles using precipitation methods on an industrial scale. therefore, the aim of this study is to analyze the economic evaluation in the project of manufacturing magnesium oxide nanoparticles using precipitation methods on an industrial scale. this evaluation is carried out from two perspectives, namely the engineering perspective and an economic perspective. 119 | international journal of informatics information system and computer engineering 1(1) (2020) 117-128 𝑡𝑟=1 from the engineering perspective, it can be determined using stoichiometric calculations and evaluation of the initial factory design. meanwhile, from the economic perspective it is determined by several parameters to determine the benefits of the project to be established, namely gross profit margin, payback period, and cumulative net present value under certain conditions (nandiyanto et al., 2018). 2. methodology in this study, selected research on the manufacture of magnesium oxide nanoparticles was conducted as the main reference (alvionita and astuti, 2017). in the economic evaluation, an analysis of the prices of equipment, utilities, and available raw materials for the manufacture of magnesium oxide nanoparticles was obtained from the online shopping site alibaba. the price of electricity per kwh is obtained based on data on electricity costs from the state electricity company. then, the data is calculated using microsoft excel with reference to several parameters, such as gross profit margin, payback period, and cumulative net present value of various cost variables. calculations were carried profitability of a project. this analysis is estimated by reducing the cost of selling the product with the cost of raw materials. gpm = 𝛴𝑡𝑟 (𝑆 . 𝜂 − 𝑅𝑀)𝑃𝐶 . 𝑄 . 𝑡 (1) s is total sales, rm is total raw materials, pc is production capacity, q is capacity of raw materials included and used in the process (kg/hour), and t is production time. ● payback period (pbp) is a calculation to predict the length of time required for an investment to return the initial capital expenditure. in short, the payback period is calculated when the cumulative net present value reaches zero. ● cumulative net present value (cnpv) is the total value of net present value (npv) from the beginning of the factory construction until the end of the factory operation. npv = 𝛴𝑡𝑟 ( 𝑅𝑡 ) (2) out based on the literature (nandiyanto et 𝑡𝑟=1 (1+𝑖)𝑡𝑟 al., 2018; ragathita et al., 2019; nassar et al., 2017; garret, 2012). to obtain the results of this study, calculations were carried out using several formulas such as: ● gross profit margin (gpm) is the first analysis to determine the level of rt is the net cash inflows minus outflows over a period of tr, i is the discount rate that can be obtained in alternative investments, tr is the project time (in a year), and tr is the last year of the project. febriani et al. computation application: techno-economic analysis on the production…| 120 3. results and discussion 3.1. engineering perspective in this study, several assumptions were used based on the illustration of the process of making magnesium oxide nanoparticles shown in fig. 2. based on these assumptions, it shows that by increasing the project through stoichiometric calculations, the production of magnesium oxide nanoparticles is approximately 1,500 kg in one cycle. the assumptions are: (1) all raw materials are upgraded to 10,000 times the laboratory scale in the literature. (2) the materials used are of high purity. (3) peg 2000, magnesium nitrate, sodium bicarbonate, and sodium hydroxide were reacted and produced magnesium oxide nanoparticles with a purity of 98%. (4) loss during the process of moving, drying, and collecting products is 2%. there are several assumptions used to ensure economic analysis. this assumption is needed to analyze and predict several possibilities that occur during the project. these assumptions are: (1) all analyzes use usd (1 usd = 14,312 rupiah); (2) based on commercially available prices, the prices of peg 2000, magnesium nitrate, sodium bicarbonate, and sodium hydroxide are 2.00 usd/kg, 2.00 usd/kg, 0.25 usd/kg, and 0 ,90 usd/kg. all materials are estimated based on stoichiometric calculations; (3) when project land has been purchased, land costs are added at the beginning year of the factory construction and recovered at the end of the project; (4) total investment cost (tic) is calculated based on lang factor (garret, 2012); (5) total investment cost is prepared in at least two steps. the first step is 40% in the first year and the second step is the remainder (during project development); (6) depreciation is estimated using direct calculation [15]; (7) one cycle of magnesium oxide nanoparticle manufacturing process takes 7.1 hours; (8) magnesium oxide nanoparticles are priced at 2 usd/pack (1 kg); (9) a one-year project is 300 days (and the remaining days are used to clean and organize the process); (10) to simplify utility, utility units can be described and converted into electrical units, such as kwh. then, the unit of electricity is converted into charge. the unit of electricity (kwh) is multiplied by the cost of electricity. the assumed annual utility cost is 41,271 usd/kwh; (12) total salary/labor is assumed to be at a fixed value of 10 usd/day; (13) discount rate 15% per year; (14) income tax is 10% per year; (15) the duration of the project operation is 10 years. economic evaluation is carried out to test the feasibility of a project. this economic evaluation is done by varying the value of raw materials, utilities, sales, salaries, and taxes under several conditions. variations in raw materials, utilities, sales, and salaries are carried out at 50, 75, 100, 125, 150, 175, and 200%, while tax variations are carried out at 10, 25, 50, 75, and 100%. 121 | international journal of informatics information system and computer engineering 1(1) (2020) 117-128 fig. 2. illustration of a flow chart for the manufacture of magnesium oxide nanoparticles table 1. table of process flow diagrams for the manufacture of magnesium oxide nanoparticles. no symbol information 1 r-1 reactor-1 2 r-2 reactor-2 3 r-3 reactor-3 4 pu-1 pump-1 5 pu-2 pump-2 6 s-1 storage-1 7 s-2 storage-2 8 fi-1 filtration-1 9 w-1 washer-1 10 o-1 oven-1 11 fu-1 furnace-1 fig. 2 shows the process of making magnesium oxide nanoparticles using the precipitation method based on the literature study (alvionita and astuti, 2017). all symbols are shown in fig. 2 are presented in table 1. mgo nanoparticles were synthesized using the precipitation method with the addition of peg, namely peg 2000 1 m mgno3 was mixed with peg and then stirred in the reactor for 10 minutes at room temperature. next, 1 m nahco3 solution was added to the previously mixed mgno3, and peg was stirred at febriani et al. computation application: techno-economic analysis on the production…| 122 a constant speed 1 m naoh solution was added to the previous solution mixture and stirred in the reactor for 3 hours without changing any parameters. from this process, a white precipitate of magnesium oxide is produced. after that, the white precipitated powder of magnesium oxide was filtered using a filtration machine. the filter results were washed with double distilled water to make a precipitate free from foreign elements and produce a substrate of mg(oh)2. then mg(oh)2 was dried using an oven for 1 hour at 80oc and sintered for 3 hours at 600oc using a furnace (alvionita and astuti, 2017; meenakshi et al., 2012). in this process, one cycle produces 11,250 kg of magnesium oxide nanoparticles. in one month, the project can produce 281,250 kg, and in one year, the project can produce 3,375,000 kg magnesium oxide nanoparticles. from an engineering point of view, the total cost for purchasing raw materials for one year is 483,750 usd, while the total sales in one year are 2,025,000,000 usd. the profit for one year is 1,881,184,752 usd. the price for the equipment cost analysis is 44,943 usd. total investment cost must be less than 190,558 usd. project life is 10 years. in 10 years, it produces magnesium oxide nanoparticles with a cumulative net present value/total investment cost reaching 30,140.28%. in the tenth year and the third year, the payback period has been successfully achieved. 3.2. economic evaluation ideal condition the ideal condition from an economic perspective is shown in fig. 3. fig. 3 shows a graph of the relationship between cumulative net present value/total investment cost over 10 years. the y-axis is the cumulative net present value/total investment cost, and the x-axis is the lifetime (years). the curve shows that there is a negative value from cumulative net present value/total investment cost (%), which is a value below 0 in the first year to the third year which indicates a decrease in income in that year due to initial capital costs for the production of magnesium oxide nanoparticles. in the third year, there is an increase in the movement of the curve that indicates an increase in income; this condition is called the payback period (pbp). profits can cover the initial capital that has been issued and continues to increase thereafter until the tenth year. in table 2 cumulative net present value/total investment cost, there is a negative value from the first year to the second year. then the value of cumulative net present value/total investment cost began to return to a positive value in the third year. thus, the production of magnesium oxide nanoparticles can be considered a profitable project because it requires a short time to recover the investment costs. 123 | international journal of informatics information system and computer engineering 1(1) (2020) 117-128 fig. 3. ideal conditions for net present value/total investment cost for life time (years) table 2. table of process flow diagrams for the manufacture of magnesium oxide nanoparticles. cnpv/tic year 0 0 -0,409351928 1 -0,845204551 2 5841,0101357171 3 10920,8843446460 4 15338,1662654538 5 19179,2809791997 6 22519,3807302830 7 25423,8152964425 8 30145,5803756431 9 30145,5803756431 10 febriani et al. computation application: techno-economic analysis on the production…| 124 effect of external conditions the success of a project can be influenced by external factors. one of the factors is the taxes levied on projects by the state to finance various public expenditures. fig. 4 shows a graph of cumulative net present value with various tax variations for ten years. the y-axis is cumulative net present value/total investment cost (%), and the x-axis is the lifetime (years). fig. 4 shows that the conditions from the beginning of the year to the second year show the same results because the cumulative net present value is under tax variations and due to project development. in addition, from the beginning of the year until the second year, there was no income, and there was a decrease by the ideal condition graph. profits continue to increase after reaching the payback period (pbp) up to the tenth year. cumulative net present value/total investment cost in the tenth year for each variation of 10, 25, 50, 75, and 100% is 301.46; 251.21; 167.47; 83.72, and -0.01%. therefore, the maximum tax to earn a break-even point (the point where there is profit and loss in the project) is 75%. changes in tax up to more than 75%, which lead to failure in the project. fig. 4. cumulative net present value curve of tax variations. change in sales the graph of the relationship between cumulative net present value and various sales variations is shown in fig. 5. the y-axis is cumulative net present value/total investment cost, and the x axis is the lifetime (years). fig. 5 is shown the results of the payback period. the conditions from the beginning of the year to the second year of the project, the cumulative net present value in various variations are the same. this happened because of the project development. the effect of sales on cumulative net present value can be obtained after the project is made for two years from the initial conditions. the greater the sales value, the more profits will be obtained from the project being implemented. however, if there are conditions that cause product sales to decline, then the project's profits will fall from ideal conditions. based on payback period (pbp) analysis, the payback period for sales variations of 50, 75, 100, 125, 150, 175, and 200% can be achieved in the third year. profit continues to increase after reaching the payback period. the profit margin generated every year increases as sales increase from ideal conditions. cumulative net present value/total investment cost in the tenth year for each variation of 50, 75, 100, 125, 150, 175, and 200% is 15055, 76; 22600.67; 30145.58; 37690.49; 45235.40; 52780.31; and 60325.22%. the minimum sales to get the break-even point (the point at which the project's profit or loss) is 50%, so the sale of magnesium oxide nanoparticles will be more profitable if the sales increase by more than 50% because the graph shows a higher cumulative net present 125 | international journal of informatics information system and computer engineering 1(1) (2020) 117-128 value/total investment cost positive value, it means the project is feasible. fig. 5. cumulative net present value curve of sales variation. variable cost changes (raw material costs, salaries, utilities) the condition of the cost of raw materials, salaries, and utilities are some of the internal factors that can affect the success of a project. fig. 6 shows a graph of cumulative net present value with various variable costs of raw materials for 10 years. the y-axis is the cumulative net present value/total investment cost, and the x-axis is the lifetime (years). the analysis is carried out by decreasing and increasing the cost of raw materials by 25, 50, 75, and 100%. the ideal raw material cost is 100%. when 25 and 50% reduce the cost of raw materials, the cost of raw materials becomes 75 and 50%, respectively. when 25, 50, 75, and 100% increase the raw material cost, the raw material cost will be 125, 150, 175, and 200%. the payback period is obtained from the variable cost of raw materials. the results of the payback period are shown in fig. 6. the conditions from the beginning of the year to the second year of the cumulative net present value project at various raw material variable costs are the same. this is due to the development of the project. the effect of raw material costs on the cumulative net present value can be obtained after the project is made for two years from the initial conditions. the lower the cost of raw materials, the higher the project profit will be. however, if there are circumstances that cause the cost of raw materials to increase, then the project's profits will fall from the ideal state. based on payback period analysis, profits continue to increase after reaching the payback period (pbp) until the tenth year. however, the profit margin earned every year is getting smaller in line with the increase in raw material costs from ideal conditions. on the other hand, the annual profit margin increases with the decrease in raw material costs from ideal conditions. cumulative net present value/total investment cost in the tenth year for each variation of 50, 75, 100, 125, 150, 175, and 200% is 30149.46; 30147.52; 30145.58; 30143.64; 30141.70; 30139.77; and 30137.83%. from the variable costs of raw materials, the project can still run and generate profits. fig. 6. cumulative net present value curve of sales variation. fig. 7 shows a graph of cumulative net present value with various salary variations. the y-axis is the cumulative net present value/total investment cost, and the x-axis is the lifetime (years). febriani et al. computation application: techno-economic analysis on the production…| 126 increasing and decreasing salaries carried out the analysis by 25, 50, 75, and 100%. the ideal salary is 100%. when 25 and 50% reduce the salary, the salary will be 75 and 50%, respectively. when 25, 50, 75, and 100% increase the salary, the salary becomes 125, 150, 175, and 200%, respectively. the payback period is obtained from the results of salary variations. the results of the payback period are shown in fig. 7. the conditions from the beginning of the year to the second year of the cumulative net present value project from various salary variations are the same. it happened because of the project development. the effect of salary on cumulative net present value can be obtained after the project is made for two years from the initial conditions. there is no significant change from the salary variation curve to the cumulative net present value graph. the payback period for each variation in salary is still achieved in the third year. however, the cumulative net present value/total investment cost in the tenth year is different for each variation. the difference in values for each variation of 50, 75, 100, 125, 150, 175, and 200% is 30157.62; 30151.60; 30145.58; 30139.56; 30133.54; 30127.52; and 30121.49%. from the salary variations, it can be concluded that the project can still run and generate profits. fig. 7. cumulative net present value curve of salary variation. fig. 8 shows the cumulative net present value graph with various utility variations. the y-axis is the cumulative net present value/total investment cost, and the x-axis is the lifetime (years). the analysis is done by increasing and decreasing the utility by 25, 50, 75, and 100%. the ideal utility is 100%, when 25 and 50% reduce the utility; the utility becomes 75 and 50%, respectively. when 25, 50, 75, and 100% increase utility, the utility becomes 125, 150, 175, and 200%, respectively. the payback period is obtained from the results of utility variations. the results of the payback period are shown in fig. 8. the conditions from the beginning of the year to the second year from the cumulative net present value to the cumulative net present value can be obtained after the project is made for two years from the initial conditions. from the utility variation to the cumulative net present value graph, there is no significant change. however, the cumulative net present value/total investment cost differs in the tenth year in each variation. the difference in values for each variation of 50, 75, 100, 125, 150, 175, and 200% is 30145.91; 30145.75; 30145.58; 30145.42; 30145.25; 30145.08; and 301344.92%. the payback period for each utility variation is still achieved in the third year. from the utility variation, it can be concluded that the project can still run and generate profits. 127 | international journal of informatics information system and computer engineering 1(1) (2020) 117-128 fig. 8. cumulative net present value curve of utility variations. 4. conslusion based on the techno-economic analysis that has been carried out, the project of references manufacturing magnesium oxide nanoparticles from an engineering point of view shows that scale-up of the project can be carried out using currently available tools and relatively low cost. the payback period analysis shows that the investment will experience a profit after more than three years. this happens because the use of raw materials in the synthesis process of magnesium oxide nanoparticles with the precipitation method is cheap and requires a short time to produce magnesium oxide. from this economic evaluation analysis, it can be concluded that this project is feasible to run. agrawal, r. m., charpe, s. d., raghuwanshi, f. c., & lamdhade, g. t. (2015). synthesis and characterization of magnesium oxide nanoparticles with 1: 1 molar ratio via liquid-phase method. international journal of application or innovation in engineering & management, 4(2), 141-145. alvionita, n., & astuti, a. (2017). sintesis nanopartikel magnesium oksida (mgo) dengan metode presipitasi. jurnal fisika unand, 6(1), 89-92. balakrishnan, g., velavan, r., batoo, k. m., & raslan, e. h. (2020). microstructure, optical and photocatalytic properties of mgo nanoparticles. results in physics, 16, 103013. ding, y., zhang, g., wu, h., hai, b., wang, l., & qian, y. (2001). nanoscale magnesium hydroxide and magnesium oxide powders: control over size, shape, and structure via hydrothermal synthesis. chemistry of materials, 13(2), 435-440. essien, e. r., atasie, v. n., okeafor, a. o., & nwude, d. o. (2020). biogenic synthesis of magnesium oxide nanoparticles using manihot esculenta (crantz) leaf extract. international nano letters, 10(1), 43-48. garrett, d. e. (2012). potash: deposits, processing, properties and uses. springer science & business media. kantam, m. l., pal, u., sreedhar, b., & choudary, b. e. (2007). an efficient synthesis of organic carbonates using nanocrystalline magnesium oxide. advanced synthesis & catalysis, 349(10), 1671-1675. febriani et al. computation application: techno-economic analysis on the production…| 128 meenakshi, s. d., rajarajan, m., rajendran, s., kennedy, z. r., & brindha, g. (2012). synthesis and characterization of magnesium oxide nanoparticles. elixir nanotechnology, 50, 10618-10620. nandiyanto, a. b. d. (2018). cost analysis and economic evaluation for the fabrication of activated carbon and silica particles from rice straw waste. journal of engineering science and technology, 13(6), 1523-1539. nassar, m. y., mohamed, t. y., ahmed, i. s., & samir, i. (2017). mgo nanostructure via a sol-gel combustion synthesis method using different fuels: an efficient nano-adsorbent for the removal of some anionic textile dyes. journal of molecular liquids, 225, 730-740. ragadhita, r. i. s. t. i., nandiyanto, a. b. d., maulana, a. c., oktiani, r. o. s. i., sukmafitri, a. j. e. n. g., machmud, a. m. i. r., & surachman, e. (2019). techno-economic analysis for the production of titanium dioxide nanoparticle produced by liquid-phase synthesis method. journal of engineering science and technology, 14(3), 1639-1652. taghavi fardood, s., ramazani, a., & woo joo, s. (2018). eco-friendly synthesis of magnesium oxide nanoparticles using arabic gum. journal of applied chemical research, 12(1), 8-15. tai, c. y., tai, c. t., chang, m. h., & liu, h. s. (2007). synthesis of magnesium hydroxide and oxide nanoparticles using a spinning disk reactor. industrial & engineering chemistry research, 46(17), 5536-5541. yunita, f. e., natasha, n. c., sulistiyono, e., rhamdani, a. r., hadinata, a., & yustanti, e. (2020, june). time and amplitude effect on nano magnesium oxide synthesis from bittern using sonochemical process. in iop conference series: materials science and engineering (vol. 858, no. 1, p. 012045). iop publishing. zhang, z., zheng, y., chen, j., zhang, q., ni, y., & liang, x. (2007). facile synthesis of monodisperse magnesium oxide microspheres via seed‐induced precipitation and their applications in high‐performance liquid chromatography. advanced functional materials, 17(14), 2447-2454. zhang, h., hu, j., xie, j., wang, s., & cao, y. (2019). a solid-state chemical method for synthesizing mgo nanoparticles with superior adsorption properties. rsc advances, 9(4), 2011-2017. 83 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 concrete mix design using particle packing method: literature review, analysis, and computation shreya sunil tolmatti*, sanskruti jaywant jadhav**, sakshi satish jadhav***, mayur m. maske**** rajarambapu institute of technology, india e-mail: *shreyatolmatti07@gmail.com a b s t r a c t s a r t i c l e i n f o particle packing technology is used to reduce the amount of cement in concrete by optimizing the concrete mix, resulting in more sustainable concrete. in this study, four different methods were used to determine the distribution of the mixture presented; packing density method, packing density method, is code method, and packing density method. the purpose of this study is to explain literature review, analysis, and data computation of the concrete mix design using particle packing method. in the packing density method, the paste content that exceeds the voids will increase along with the increase in the quality of the concrete. in cases of packing density, the cement-water ratio decreases with the quality of the concrete. in the packing of too many trials, trials and tribulations should be carried out to achieve the ratio of water-cement and paste content for a certain grade of concrete. this correlation curve helps reduce the experiments involved in determining the ratio of semen and paste content for a given concrete quality. the water and cement contents for the packing density and the is code method are almost the same for each particular concrete class. the workability of concrete achieved was more in the packing density method than the is code method for the same concrete quality, because the water-cement ratio was slightly higher in the packing density method than the is code method. therefore, more water and cement are required in terms of packing density. the correlation curve can be used to determine the ratio of water-cement and paste the content that exceeds the voids for a certain concrete quality. article history: ___________________ keywords: concrete mix design, mixtures presented, packing density method, is code method international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(1) (2021) 83-102 received 18 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 tolmatti et al. concrete mix design using particle …| 84 1. introduction there are various methods of proportioning for various types of concrete. the packing density method of mix design is the only mix design method used for proportioning normal concrete, high strength concrete, no-fines concrete, and self-compacting concrete (raj et al., 2014). the subject of optimizing the concrete composition by selecting the right amounts of various particles has already used interest for more than a century. to optimize the particle packing density of concrete, the particles should be selected to fill up the voids between large particles with smaller particles and so on, to obtain a dense and stiff particle structure. most of the early researchers, working on the packing of aggregates, proposed methods to design an ideal particle size distribution. geometrically based particle packing models can help to predict the water demand of concrete, and thus the material properties. the cement paste has to fill up the voids between aggregate particles and the “excess” paste will then disperse the aggregate particles to produce a thin coating of paste surrounding each aggregate for lubricating the concrete mix. in general, the higher the packing density of the aggregate, the smaller will be the volume of voids to be filled and the larger will be the amount of paste in excess of void for lubrication (yang et al., 2020). the purpose of this study is to explain literature review, analysis, and data computation of the concrete mix design using particle packing method. in is code method of mix design, we have curves to decide the water-cement ratio whereas in the packing density method we don’t have such type of co-relation curves available. here an attempt has been made to develop co-relation curves between compressive strength of concrete versus water-cement ratio and paste content versus compressive strength. these co-relation curves help to reduce the trials and decide the watercement ratio and paste content for the given grade of concrete. packing density is a new kind of mix design method used to design different types of concrete (glavind, m., & pedersen, e. 1999). to optimize the particle packing density of concrete, the particles should be selected to fill up the voids between large particles with smaller particles and so on, to obtain a dense and stiff particle structure. the higher degree of particle packing leads to minimum voids, maximum density, and requirement of cement and water will be less. in this work, the co-relation curves are developed for the packing density method between compression strength and water-cement ratio, paste content to reduce the time involved in the trial to decide water-cement ratio, and paste content for a particular grade of concrete (rao, v. k., & krishnamoothy, s. 1993). an important property of multiparticle systems is the packing density. this is defined as the volume fraction of the system occupied by solids. for a given population of grains, it is well known that the packing density (powers, 1968). the packing density of the aggregate mixture is defined as the solid volume in a unit total volume (hettiarachchi, c., & mampearachchi, w. k. 2020). the aim of obtaining packing density is to combine aggregate particles to minimize the 85 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 porosity, which allows the use of the least possible amount of binder (see fig. 1). fig. 1. prediction of particle packing density (source: http://link.springer.com/ accessed on 04/04/2021) 2. literature review wong and kwan used the ordinary portland cement complying with bs 12:1996. fennis and walraven used ordinary portland cement and blast furnace slag cement. wong and kwan used aggregate particles smaller than 1.2mm for mortar and aggregate particles larger than 1.2mm for concrete mix. kwan and wong used pulverized fly ash as cementations material complying with bs 3892: part 1: 1982. kwan and wong used the condensed silica fume complying with astm c 1240-03 as the cementation’s material in their experiments. kwan and wong in their studies used two types of superplasticizers a polycarboxylatebased and cross-linked polymer and naphthalene-based formaldehyde condensate 1. kwan and wong measured the packing densities of cementations materials containing ordinary portland cement, pulverized fly ash, and condensed silica fume. the results for non-blended materials revealed that the addition of a superplasticizer would always increase the packing densities of ordinary portland cement and pulverized fly ash, the addition of a polycarboxylatebased superplasticizer could decrease the packing density of condensed silica fume. 2. kwan and wong proposed a three-tier system design. the mix design would be divided into three stages. at the first stage, the packing density of the cementitious materials would determine the water demand, and at the second stage the aggregate particles smaller than 1.2mm would determine the paste demand and at the third stage, the aggregate particles larger than 1.2mm would determine the mortar demand. 3. kwan and wong used the minislump cone test to check the fresh state properties in their experimental studies. fennis and walraven carried out centrifugal consolidation to check the workability. kwan and wong obtained a curve between voids ratio and watercement ratio for cementitious materials, where the ordinary portland cement is blended with the pulverized fuel ash and condensed silica fume in different proportions. from the above study, it is observed that the packing density mix design method is used to minimize voids to increase particle packing and to reduce the binder content. 3. method ordinary portland cement conforming to is 12269-1987 locally available river sand belonging to zone ii of is 383-1970was used. the locally available crushed aggregate of size 12.5 mm and 20 mm downsize conforming to is 383-1970 were used in the preparation of concrete. potable water was used in the tolmatti et al. concrete mix design using particle …| 86 present investigation for both casting and curing of the concrete. superplasticizer complies with is 9103:1999 sulphonated napthelene based polymers are used. bulk density and specific gravity test were carried out as per is 2386(part iii)1963 and the test results are presented in table 1. 4. results and discussion 4.1. design of concrete mix using packing density method 1. determination of aggregate fractions the packing density of the aggregate mixture is defined as the solid volume in a unit total volume. the aim of obtaining packing density is to combine aggregate particles to minimize the porosity, which allows the use of the least possible amount of binder. two size fractions of coarse aggregates were selected for the study i.e., 20 and 12.5 mm down a size. the values of bulk density of the coarse aggregates (20 and 12.5 mm in size) were first determined separately. the coarse aggregate 20 and 12.5 mm were mixed in different proportions by mass, such as 90:10, 80:20, 70:30, and 60:40, etc., and the bulk density of each mixture is determined. the addition of a smaller size aggregate (12.5 mm downsize) increases the bulk density. however, a stage is reached when the bulk density of the coarse aggregate mixture, which instead of increasing, decreases again. the results of the bulk density of coarse aggregate fractions (20 mm and 12.5 mm) are plotted in fig. 1. 2. determination of packing density the packing density of individual aggregate in a volume fraction of total aggregate or overall aggregate is determined from its maximum bulk density of the mixture and specific gravity. therefore, the total packing density of the mixture is the sum of the packing density of 20 mm, 12.5 mm, and fine aggregate i.e., equal to the ratio of bulk density of mixture to the specific gravity of individual aggregate (20 mm: 12.5 mm: fine aggregate). the value of specific gravity should be taken as average if the values are differing in the third decimal and if the values are differing in the second decimal, the individual values should be taken for calculating packing density and voids content. 3. determination of voids contents and voids ratio the voids content in percentage volume of aggregate or mixture of three aggregate is determined from its bulk density. from figures 2, 3, and 4, it is observed that the bulk density, packing density are maximum and voids ratio is minimum for 70 % of coarse aggregate (20 mm) and 30 % of coarse aggregate (12.5 mm) respectively. fig. 2. maximum bulk density for 20 and 12.5 mm aggregates 87 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 fig. 3. maximum packing density for 20 and 12.5 mm aggregate fig. 4. minimum voids ratio for 20 and 12.5 mm aggregates an increase in fine aggregate particles leads to a decrease in void content thus increases the bulk density. the replacement of fine aggregates in the total coarse aggregates (20 and 12.5 mm downsize in the proportion of 70:30) in the ratio of 90:10, 80:20, 70:30, 60:40, and 55:45. by increasing the finer content the bulk density increases up to a maximum extent after which it again reduces. thus, the proportion obtained for maximum bulk density is fixed as total coarse aggregates: fine aggregates i.e., 60:40. total coarse aggregate proportion i.e., 20 mm:12.5 mm is fixed as 70:30 as mentioned earlier. therefore, proportions of these aggregates i.e., coarse aggregates 20 mm: coarse aggregates 12.5 mm: fine aggregates are 42:18:40. the bulk density, packing density, and voids ratio are plotted against the mass fraction of coarse aggregate are presented in figures 5, 6, and 7, respectively. from figures 5, 6, and 7, the maximum bulk density is 2.007 gm/cc, the maximum packing density is 0.722 gm/cc, and the minimum voids content is 0.2866. fig. 5. maximum bulk density for 20 mm and 12.5 mm and fine aggregates fig. 6. maximum packing density for 20 mm and 12.5 mm and fine aggregate fig. 7. minimum voids for 20 mm, 12.5 mm and fine aggregate tolmatti et al. concrete mix design using particle …| 88 using the above concept, design of concrete mix is carried out for m20, m25, m30, m35 and m40 concrete mixes. a detailed sample calculation for m20 grade of concrete is presented below. the ingredients of concrete for m20 grade were obtained for 5%, 10% and 15% in excess of paste content and water-cement ratio 0.56 and 0.58 the values are presented in table 2. 4.2. mix design for m20 grade concrete (packing density method) the calculations are presented in the following paragraph for bulk density, voids ratio and packing density. 1. bulk density of combined coarse aggregate 20 and 12.5 mm in the proportion 70:30. w 2 w 1 bulk density = volume of mould where, w1= empty weight of mould w2= weight of mould + aggregate filled bulk density (maximum) = 35066-9800 15000 = 1.6840 gm / cm3 2. bulk density of three aggregates i.e., ca 20mm : ca 12.5mm : fa is 42 : 18 : 40. (coarse aggregate 20 mm : 12.5 mm i.e., 70 : 30 as fixed earlier). bulk density (maximum) = 39916-9800 15000 = 2.0077 gm / cm3 3. voids content: voids content in percent volume = 2.8143-2.0077 2.8143 x 100 = 28.660% 4. packing density (p.d.): packing density (maximum) = bulk density x weight fraction specific gravity packing density of 20mm aggregates = 2.00778x0.420 2.9122 = 0.2896 gm/cm3 packing density of 12.5 mm aggregates = 2.0078x0.180 2.9376 = 0.1230 gm/cm3 packing density of fine aggregates = 2.00778x0.400 2.5931 = 0.3097 gm/cm3 total packing density = packing density of ca (20 mm) + packing density of ca (12.5 mm) + packing density of fine aggregate. pd = 0.7223 gm / cm³ this packing density value is fixed for further calculations. 4.3. determination of paste content for m20 grade concrete 89 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 minimum paste content is the sum of the void content in combined aggregate and excess paste over and above it to coat the aggregate particle. the meaning of minimum paste content can be explained as, a concrete mix containing minimum paste content should be cohesive, free from segregation and bleeding. flow table tests were carried out to decide the minimum paste contents required to form the workable mix for different w/c ratios and different paste content in excess of void content. voids content = 1 – 0.7223 = 0.2777 assuming paste content as 10% more than void content, detailed calculations to obtain all. ingredients of concrete such as coarse aggregate 20 mm, 12.5 mm, fine aggregate, cement, and water content are given below. paste content 10% more than void content paste content = 0.2777 + 0.1x 0.2777 = 0.3054 volume of aggregates = 1 – 0.3054 = 0.6945 cc total solid volume of aggregates = weight fraction of 20 mm specific gravity + weight fraction of 12.5 mm specific gravity + weight fraction of fine aggregate specific gravity total solid volume of aggregates = 0.420 2.9122 + 0.180 2.9376 + 0.400 2.5931 = 0.3598 cc weight of 20 mm aggregates = 0.6945 0.3598 x 0.420 x 1000=810.7354 kg/cum weight of 12.5 mm aggregates = 0.6945 0.3598 x 0.180 x 1000=347.4580 kg/cum weight of fine aggregates = 0.6945 0.3598 x 0.400 x 1000=722.1290 kg/cum for m20 grade concrete keeping in mind the target mean strength suitable water-cement ratio is fixed as per trial mixes. w/c ratio = 0.56; w = 0.56c. total paste=c+w= c 3.15 + 0.56c 1 =0.8775 c cement content= 0.3054 0.8775 x 1000 =348.1140 kg/cum water content = 0.56 x 348.1140 = 194.9438 kg/cum following the above procedure, all the ingredients of concrete were obtained for 5%, 10%, and 15% in excess of paste content and water-cement ratio 0.56 and 0.58, the values are presented in table 2. to decide the paste content and water cement ratio among three paste content and two water cement ratios, using the above ingredients flow table tests were carried out. flow table test is carried out as per is 1199-1959 (indian standard). results of flow table tests for m20 grade concrete indicated that water cement ratio 0.58 and all the three-paste content (i.e., 5%, 10% and 15 %) and water cement tolmatti et al. concrete mix design using particle …| 90 ratio 0.56 with 5% paste content were rejected because of segregation and bleeding. water cement ratio 0.56 with paste content of 10% and 15% in excess of void content resulted in good flow percent of 133 and 134 respectively without segregation and bleeding. for water cement ratio 0.56 in order to decide the paste content i.e., 10% and 15% in excess of void content, trial cube casting was carried out for 7 days cube compressive strength. the average compressive strength (3 cubes) obtained at the end of 7 days curing was 22.88 n/mm and 23.666 n/mm for 10% and 15% paste content respectively. keeping economy in mind paste content of 10% for water-cement ratio 0.56 was finalized for further casting. mix design is carried for m25, m30, m35, and m40 grade concrete as mentioned in mix design steps for m20 grade concrete. the value of packing density remains the same irrespective of the grade of concrete because coarse aggregate 20 mm, 12.5 mm and fine aggregate used is the same for all grades of concrete. depending on the grade of concrete paste content will vary, increases with an increase in grade of concrete. the water-cement ratio for different grades of concrete (m25, m30, m35, and m40) is fixed as per trial mixes. paste contents for different grades of concrete were determined using flow table tests as mentioned earlier for an individual grade of concrete finalized mix proportions are presented in table 3. 4.4. design of concrete mix using is code method mix design is also carried out using is code 10262-2009 (indian standard). the objective of is code method of mix design is to compare the ingredients of concrete (mix proportions) with the packing density method and also to compare the compressive strength at 28 days in these two cases and relevant observations were discussed. here also the final mix proportions were obtained for m20, m25, m30, m35, and m40 grade of concrete using is method with the different trial mixes. the trial mix design for different grades of concrete was carried for different water-cement ratios and workability is checked using flow table tests. accepted trial mixes were further used to cast the trial cube specimens and were tested for compressive strength at the 7 days curing age. observing the results of trial casting the appropriate mix is finalized. this finalized mix proportion is used for further casting. finalized mix proportions for different grades of concrete designed by is code method are presented in table 4. 4.5. comparing the mix proportion of concrete by is code method and particle packing method 1. mix proportions and compressive strength finalized mix proportions for m20, m25, m30, m35, and m40 grade concrete using packing density and is code method are presented in the following tables. using these finalized mix proportions for different grades of concrete final casting was carried out as mentioned in the following section. in the packing density method, finalized mix proportions were used for final casting. six cube specimens were cast (3 cube specimens for 7 days curing and 3 cube specimens for 28 days curing). similarly, in is method for each grade of concrete six cube specimens were cast (3 cube specimens for 7 days curing and 3 cube specimens for 28 days curing). the casting, curing, and compressive strength testing procedure was followed 91 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 according to is 516-1959 (see figs. 8 and 9). the average test result of 3 cube specimens is considered for the final test result. the results of the final casting are presented (tables 5 and 6). 2. workability the variation in workability of concrete mixes designed by different methods with different water-cement ratios is presented in table 7 (kore, s. d., & vyas, a. k. 2017). it can be seen from table 7, all the concrete mixes achieved their target slump of 75-100 mm. while achieving the target slump the dose of the superplasticizer is increased. in the case of the packing density method at all watercement ratios, the dose of superplasticizer required is more than that of the bis code method. this increase is caused by the increased sand content in concrete mixes designed by the packing density method as seen in table. on average sand, content increased by 14% in the packing density method as compared to that of the bis code method. the increased sand content absorbs more water from the mix resulted in a stiff mix. hence, a higher dose of super-plasticizers is required to achieve the desired workability. 3. saving in cement content and cost comparison table 8 saving in cement content (kore, s. d., & vyas, a. k. 2017). table 9 shows the saving in cement content in concrete mixes designed by the packing method. from this table, it was observed that the concrete mixes designed by the packing density method are economical because of saving in cement content. the maximum saving of 18% was achieved at a 0.45 water-cement ratio. the concrete mixes designed by the packing density method showed an average 12% reduction in cement content as compared to that of the bis code method. it depicts that the concrete mixes designed by the packing density method are economical (kore, s. d., & vyas, a. k. 2017). the table shows the cost analysis of the concrete mixes. from this analysis, it was observed that the concrete mixes designed by the packing density method show a saving in the material cost of concrete. the average cost of material to produce concrete can be reduced by 11% by adopting the packing density method for the design of concrete mixes. hence it indicates that the concrete mixes designed by the packing density method are cost-effective and economical. 4. co2 production the cement manufacturing industry is a major contributor to co2 emissions in the world. the contribution of the cement industry in greenhouse gas emission is around 3.95 billion tons annually and that is 7% of the total greenhouse gas emissions on the earth’s surface. the global annual production of concrete in the year 2014 was 4.2 billion tons and it is expected that this figure may increase by 2.9 % by 2018. in india, around 275 mt of cement was produced during the year 2014 which accounts for the generation of an equal amount of co2. to produce 1 ton of cement around 0.94 ton of co2 is released. the co2 emission factor for road transport, i.e., trucks or lorries is considered as 512.2 g/km. table 10 reduction in carbon di oxide emission. tolmatti et al. concrete mix design using particle …| 92 table 1. the technology used in the device of virtual voting system sl. no materials bulk density kg/m3 (compacted condition) bulk density kg/m3 (loose condition) specific gravity 1. fine aggregates 1600.133 1718.063 2.593 2. coarse aggregate 12.5 mm 1387.777 1542.222 2.937 3. coarse aggregate 20 mm 1525.555 1660.000 2.912 table 2. trial mix proportions for m20 grade concrete grade of concrete w/c ratio excess paste content (%) water content (kg/m3) cement content (kg/m3) wt. of fine aggregate (kg/m3) wt. of 12 mm coarse aggregate (kg/m3) wt. of 20 mm coarse aggregate (kg/m3) m20 0.58 5 188.4416 324.8994 787.6736 354.4531 827.0573 0.58 1 2.4243 1.0817 2.5455 0.58 10 197.4151 340.3708 772.2352 347.5058 810.8469 0.58 1 2.2688 1.0209 2.3822 0.58 15 206.3885 355.8422 756.7967 340.5585 794.6366 0.58 1 2.1268 0.9570 2.2331 0.56 5 186.0907 332.3048 787.6736 354.4531 827.0573 0.56 1 2.3703 1.0667 2.4889 0.56 10 194.9522 348.1289 772.2352 347.5058 810.8469 0.56 1 2.2182 0.9982 2.3292 0.56 15 203.8136 363.9529 756.7967 340.5585 794.6366 0.56 1 2.0794 0.9357 2.1834 93 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 table 3. finalized mix proportions designed by packing density method grade of concrete w/c ratio excess paste content (%) water content (kg/m3) cement content (kg/m3) wt. of fine aggregate (kg/m3) wt. of 12 mm coarse aggregate (kg/m3) wt. of 20 mm coarse aggregate (kg/m3) m20 0.56 10 194.9522 348.1289 772.2352 347.5058 810.8469 0.56 1 2.2182 0.9982 2.3292 m25 0.54 15 201.1187 372.4420 756.7967 340.5585 794.6366 0.54 1 2.0320 0.9144 2.1336 m30 0.50 20 203.8259 407.6518 741.3583 333.6112 778.4262 0.50 1 1.8186 0.8184 1.9095 m35 0.48 25 208.9378 435.2871 725.9199 326.6639 762.2159 0.48 1 1.6677 0.7505 1.751 m40 0.44 30 209.7061 476.6047 710.4815 319.7167 746.0055 0.44 1 1.4907 0.6708 1.5653 table 4. finalized mix proportions designed by is code method grade of concrete w/c ratio water content (kg/m3) cement content (kg/m3) wt. of fine aggregate (kg/m3) wt. of 12 mm coarse aggregate (kg/m3) wt. of 20 mm coarse aggregate (kg/m3) m20 0.55 192 349 669 609.7 609.7 0.55 1 1.9169 1.7470 1.7470 m25 0.52 192 369.23 662.786 633.70 633.70 0.52 1 1.7950 1.7162 1.7162 m30 0.48 197 410 646 617 617 0.48 1 1.5756 1.5048 1.5048 m35 0.46 197 428 638 610.5 610.5 0.46 1 1.4907 1.4264 1.4264 m40 0.42 197 469 625 598 598 0.42 1 1.3326 1.2750 1.2750 tolmatti et al. concrete mix design using particle …| 94 table 5. compressive strength of cube cast using is code method grade of concrete w/c ratio paste content strength of cube (mpa) (7 days) strength of cube (mpa) (28 days) m20 0.56 10 % 22.8889 33.7037 m25 0.54 15 % 26.9629 38.7407 m30 0.5 20 % 30.3333 44.4444 m35 0.48 25 % 36.6667 50.6667 m40 0.44 30 % 40.8519 54.8048 table 6. compressive strength of cube cast using packing density method grade of concrete w/c ratio strength of cube (mpa) (7 days) strength of cube (mpa) (28 days) m20 0.55 22.6667 31.5555 m25 0.52 26.2222 37.7037 m30 0.48 31.8518 45.6296 m35 0.46 34.0741 48.8889 m40 0.42 38.2222 54.5184 fig. 8. variation in compressive strength of concrete mixes designed by packing density method 95 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 fig. 9. variation in compressive strength of concrete mixes designed by bis code method table 7. variation in workability of concrete mixes wc bis code method packing density method dose of superplasticizer by weight of cement (%) slump (mm) dose of superplasticizer by weight of cement (%) slump (mm) 0,55 0,25 80 0,50 100 0,50 0,25 90 0,65 90 0,45 0,50 100 0,90 90 0,40 0,70 90 1,0 90 0,35 0,25 80 2,3 100 table 8. saving in cement content cement content in kg/m3 w/c ratio bis code method packing density method saving in cement by pdm (kg) % saving in cement 0,55 348 308 40 11,49 0,5 383 327 56 14,62 0,45 425 348 77 18,12 0,40 400 372 28 7,00 0,35 435 400 35 8,05 tolmatti et al. concrete mix design using particle …| 96 table 9. cost of material for production of 1m3 of concrete cost of material in rs. for 1m3 of concentrate w/c ratio bis code method packing density method saving in cost percentage saving 0,55 2817 2640 177 6 0,50 2995 2746 149 8 0,45 3217 2864 353 11 0,42 3295 2998 297 9 0,35 3311 3155 156 5 table 10. reduction in carbon di oxide emission cement content in kg/m3 co2 emission (kg/m3) percentage reduction in co2 emission by packing density method w/c ratio bis code method packing density method bis code method packing density 0,55 348 308 327 289 11 0,5 383 327 360 307 15 0,45 425 348 400 327 18 0,40 400 372 376 350 7 0,35 435 400 409 376 8 from the above table, it can be observed that the cement content required for the design of concrete mixes by using the packing density method reduced by 12%. on the other hand, from table 8, it can be observed that the reduced usage of cement content for concrete mix resulted in an average 12% reduction in carbon dioxide emission as compared to that of the bis code method. adopting a packing density approach for the design of concrete mixes would reduce the annual global cement production from 4.2 billion tons by 0.51 billion tons and co2 release from 3.95 billion tons to 3.47 billion tons. the concrete produced using the packing density approach is not only a costeffective and sustainable product 97 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 mitigating environmental pollution to a large extent. 4.6. applications of the mixture design approach using particle packing model to demonstrate the application of the particle packing models to the mixture design of various concretes, the software lisa 20 is considered because of its relevance, simplicity, and availability. a detailed description of the method of a mixed design using this software is given in the user manual of lisa. for this study, the modified andreassen model was chosen. 1. application in mortar to demonstrate the modified andreassen model for the concrete application, a pilot trial on cement mortar was conducted. the proportions used for the testing of compressive strength of cement mortar as per is 4031: 1996 21 were used to design the reference mix and denoted as mix a. then the missing zones of particles in the particle gradation were adjusted using crushed sand, quartz powder, and micro silica and denoted as mix b. the mixture design details are given in table 3. the flow table spread diameter value at 25 blows was kept constant at 60 percent by adding polycarboxylic ether (pce) based admixture. the appearance of the ideal gradation curve against the actual overall gradation curve for mix a and mix b is shown in fig 5. it can be seen in fig 5 that the reference gradation curve (smooth curve) is the modified andreassen curve, and the actual overall particle size distribution curve is the irregular curve. this is adjusted to fit the reference curve to the closest extent possible by altering the inputs by trial and error. the results of the compressive strength of the two mortar mixtures are shown in fig 7. the value shown is the average strength of three 7 cm cube specimens. it may be noted that from table 3, the w/c of 0.4 is kept constant for both mix a and mix b; only the gradation of mix b is adjusted to fit the andreassen curve. it is observed from fig 6 that a strength increase of about 28-30 percent could be achieved at all ages for mix b. a comparison of figs 5 (a) and (b) and the examination of the increase in compressive strengths, fig 6, due to the altered proportions of the mortar by filling the missing fractions showed that the model by and reason may lead to optimal mixture proportions (see fig. 10 and table 11). 2. application in concrete the mixture designs for high strength, high performance, and selfcompacting concrete are illustrated below. • high strength concrete (hsc) high-strength concrete could be designed at low cement content with a proper selection of ingredients. a typical mixture proportion of high-strength concrete is given in column 2 of table 4. the modified andreassen ideal gradation curve for q = 0.26 and the actual overall gradation are shown in fig 7(a). • high-performance concrete (hpc) high-performance concrete requires the usage of supplementary cementations materials. a typical mixture proportioning using micro silica with the aid of the modified andreassen model with exponent q = 0.27 and the combined gradation that could be managed with given materials are shown in fig 7(b). the tolmatti et al. concrete mix design using particle …| 98 details of the mixture proportions are given in column 3 of table 4. • self-compacting concrete (scc) self-compacting concrete is highly flowable and is very sensitive to overall gradation and water content. a typical mixture proportion of scc is given in column 4 of table 4. the ideal gradation for q = 0.22 and the combined grading obtained with available material are shown in fig 7(c) (see fig. 11 and table 12). table 11. mixture design details of mortar mixtures used for validation of the model ingredients, kg/m3 mix a mix b standard sand (g-1) 541 341 standard sand (g-2) 541 341 standard sand (g-3) 541 341 crushed sand (correction) 341 cement, (opc) 541 541 quartz powder (insert filler) 150 micro silica 60 water 216 216 sp (pce) 2 4 w/c 0,40 0,40 w/p 0,40 0,29 flow spread, percent 60 60 fig. 10. comparison of strength between mix a and mix b 99 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 fig. 11. ideal grading curve and actual overall particle size distribution for (a) high strength concrete (b) high-performance concrete (c) self-compacting concrete tolmatti et al. concrete mix design using particle …| 100 table 12. proportions of various particulate ingredients ingredients, kg/m3 hsc hpc scc coarse aggregate (<20 mm) 565 520 coarse aggregate (<10 mm) 545 530 748 fine aggregate (<4,75 mm) 900 868 870 cement (opc 53 grade) 270 360 320 inert filler (quartz) 55 fly ash 220 micro silica 30 42 water 12 144 180 superplasticizer, pce, / 5 8.25(snf)2.12 2.12 viscosity modifier. / 0,375 exponent (q) 0,26 0,27 0,22 workability – slump, mm 100 100 690 (flow) test result compressive strength, mpa 3-day 42 47 14 7-day 63 63 21 28-day 83 78 41 rcpt coulombs 350 1296 hsc: high strength concentrate, hpc: high performance concentrate, scg: self compacting concentrate, rcpt: rapid chloride permeability test. 5. conclusion the packing density value will remain the same irrespective of the grade of concrete. in the packing density method, paste content more than void content will increase with the increase in the grade of concrete. in the case of the packing density method, the watercement ratio decreases with an increase in the grade of concrete. in packing density too many trial calculations, trial tests, and trial casting are to be done to arrive at water-cement ratio and paste content for 101 | international journal of informatics information system and computer engineering 2(1) (2021) 83-102 a particular grade of concrete. these corelation curves help to reduce the trials involved in determining the watercement ratio and paste content for the given grade of concrete. the water and cement content for packing density and is code method is nearly the same for any grade of concrete. the workability of concrete achieved is more in packing density method compared to is code method for the same grade of concrete, as the water-cement ratio is slightly higher in packing density method than is code method. the fine aggregate particles required are more in the case of the packing density method compared to is code method. therefore, water and cement required in case of packing density are more. co-relation curves can be used to decide the water-cement ratio and paste content more than void content for the given grade of concrete. references glavind, m., & pedersen, e. (1999). packing calcuations applied for concrete mix design. utilizing ready mix concrete and mortar, thomas telford publishing, 121-130. hettiarachchi, c., & mampearachchi, w. k. (2020). effect of surface texture, size ratio and large particle volume fraction on packing density of binary spherical mixtures. granular matter, 22(1), 1-13. indian standard code of practice for, “concrete admixtures – specification”, is 9103 – 1999, bureau of indian standards, new delhi. indian standard code of practice for, “methods of sampling and analysis of concrete”, is 1199 – 1959, bureau of indian standards, new delhi indian standard code of practice for, “specification for 53 grade ordinary portland cement”, is 12269 – 1987, bureau of indian standards, new delhi. indian standard code of practice for, “specification for coarse and fine aggregates from natural sources for concrete”, is 383 – 1970, bureau of indian standards, new delhi. kore, s. d., & vyas, a. k. (2017). packing density approach for sustainable development of concrete. journal of materials and engineering structures «jmes», 4(4), 171-179. powers, t. c. (1968). properties of fresh concrete, john wiley and sons. inc., new york, 301. raj, n., patil, s. g., & bhattacharjee, b. (2014). concrete mix design by packing density method. iosr journal of mechanical and civil engineering, 11(2), 34-46. tolmatti et al. concrete mix design using particle …| 102 rao, v. k., & krishnamoothy, s. (1993). aggregate mixtures for least-void content for use in polymer concrete. cement, concrete and aggregates, 15(2), 97-107. wong, h. h., & kwan, a. k. (2008). packing density of cementitious materials: part 1—measurement using a wet packing method. materials and structures, 41(4), 689-701. yang, y., chen, b., su, y., chen, q., li, z., guo, w., & wang, h. (2020). concrete mix design for completely recycled fine aggregate by modified packing density method. materials, 13(16), 3535. 1 | international journal of informatics information system and computer engineering 1 (2020) 1-12 modeling traffic flows with fluid flow model paulus setiawan suryadjaja *, maclaurin hutagalung **, herman yoseph sutarto*** department of information technology, institut teknologi harapan bangsa, jl. dipatiukur no. 80-84, bandung, indonesia e-mail: *paulus1945s@gmail.com **maclaurin@ithb.ac.id ***hytotok@gmail.com a b s t r a c t s a r t i c l e i n f o this research presents a macroscopic model of traffic flow as the basis for making intelligent transportation system (its). the data used for modeling is the number of passing vehicles per three minutes. the traffic flow model created in the form of fluid flow model (ffm). the parameters in the model are obtained by mixture gaussian distribution approach. the distribution consists of two gaussian distributions, each representing the mode of traffic flow. in the distribution, intermode shifting process is illustrated by the first-order markov chain process. the parameters values are estimated using the expectationmaximization (em) algorithm. after the required parameter values are obtained, traffic flow is estimated using the observation and transition-based most likely estimates tracking particle filter (otpf). to examine the accuracy of the model has been made, the model estimation results are compared with the actual traffic flow data. traffic flow data is collected on monday 20 september 2017 at 06.00 to 10.00 on dipatiukur road, bandung. the proposed model has accuracy with mape value below 10% or falls into highly accurate categories. article history: received 1 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 ___________________ keywords: fluid flow model, expectation-maximization, particle filter, otpf, intelligent transportation systems, gaussian distributions, mixture gaussian, markov chain. 1. introduction traffic jam in bandung city has been at an alarming rate. the survey results of international journal of informatics information system and computer engineering journal homepage: http://ejournal.upi.edu/index.php/ijost/ international journal of informatics information system and computer engineering 1 (2020) 1-12 suryadjaja et al. modeling traffic flows with fluid flow model | 2 the ministry of transportation showed that bandung is the seventh most congested city in indonesia. the average vehicle speed in the city of bandung is only 14.3 kilometers per hour and the value of volume to capacity (vc) ratio streets in the city of bandung for 0.85 (wiyono, a, s., 2014). for that, it needs to be applied more efficient traffic engineering system by utilizing modern technology, for example intelligent transportation system (its). its is a utilization of information technology elements such as sensors, traffic models, software, and telecommunications networks to improve the security and efficiency of traffic systems. its was created to address the common problems of transportation systems that are less efficient, including air pollution in the streets, fuel wastage, transportation security, and waste of time due to traffic congestion (ali, m., 2011). the control system that it applies requires an accurate model describing the behavior of traffic on the road network and the intersection for optimal performance. the model that it takes its system is a model of traffic flows and a long model of queuing at the intersection (sutarto, y, h., 2016). the behavior of traffic in an area is not the same as other regions, so it is necessary to create a model separately for each region that applies its. if the city of bandung wants to implement its effective, it needs to be made accurate traffic model based on traffic behavior in bandung city. the traffic flow model created in this study is a model of the traffic flow with a macroscopic scale in the form of the fluid flow model (ffm). ffm has the advantage of other macroscopic traffic flow models in terms of the ease of gathering and processing of traffic flow data. as a comparison, the model of the spatio-temporal random effects (stre) , which was designed on research (yaojan., et al, 2012) requires a large number of traffic flow data, i.e. two-week traffic flow data to be able to produce satisfactory estimation. in addition, the model can produce an estimate with a high accuracy of mape 8% to 15% for traffic flow data on the main road between cities. as for traffic flow data in the center of the crowd, its accuracy is dropped to the mape number gained to 20% to 29.6%. the platoon-based traffic flow model created in research (marinica, n., and boel, r., 2012) can be used to efficiently regulate the light length of red and green lights at the traffic lights. however, because the model modeled the movement of a group of vehicles, traffic flow data and the capacity of the road segments and the intersection are interconnected on the modelled street area. traffic flow models can also be created based on the queue theory, as is done in research (vandaele, n., et al., 2000). however, the steps that need to be done are more complicated. there are three variables that need to be collected data for later estimated namely the arrival rate (arrival rate), the service rate (service rate), and the maximum density (maximum traffic density). once these three variables are estimated, the maximum traffic flow values that can pass the modelled roads can be calculated. the formulation of the problem with the estimation process of the parameter and state is described in part ii of this paper. part iii explains the process of testing the accuracy of models that have been made using traffic flow data on dipatiukur road, bandung. part iv contains the 3 | international journal of informatics information system and computer engineering 1 (2020) 1-12 conclusion of this study as well as further research advice on development in the field of traffic flow modeling. 2. main content this section describes the process of creating a ffm traffic flow model. the process of creating models in the research consists of several stages namely the formulation of problems, estimation of parameters, and estimates state. 2.1. problem formulation fluid flow model (ffm) analyses the flow of traffic as fluid flow. ffm described the flow with a random variable which indicates the number of vehicles crossing a section tk y of a road at a time interval tk , tk 1 . mathematically, ytk can be defined as: tk y ( ) 1+ − tk k k n t t (1) with ntk represents the number of vehicles passing through the observation point at the time interval tk , tk 1 . based on that definition, the value of ytk is a nonnegative number real. because the definition states that at the time interval tk , tk 1 there is a vehicle as much as ytk that crosses the observation point, so implicitly there is an assumption that the traffic flow at the time interval tk , tk 1 of unchanged (constant) and distance the vehicles passing through at that time interval were uniform (sutarto, y, h., 2016), (sutarto, y, h., and joelianto, e., 2015). in this study, to form the ffm, the random variable ytk was described with the mixed gaussian distribution approach (mixture gaussian). gaussian distribution of the mixture is a system that is illustrated with some modes which are each a gaussian distribution. on that distribution, the intermode is set with the process markov chain first order. in this research, traffic flow is categorized into two modes. the first mode represents normal traffic flow with a lower ytk value and a second mode that represents crowded traffic with a higher ytk value. with the gaussian distribution of such mixtures, it is expected that the traffic flow can be described by mode which each has different characteristics (sutarto, y, h., and joelianto, e., 2015). the model is made to form a nongaussian system, since the mixed gaussian distribution not only contains a gaussian distribution, but also contains a markov chain component. each mode traffic flow is a random variable with a gaussian distribution. thus, the individual mode has gaussian distribution parameters, i.e. average and variance. to explain the inter-mode displacement process, the model also needs to contain a markov chain component. the markov chain (markov chain) is used to estimate conditions in the present time based on known past conditions. the markov chain process is generally described with transition probability matrix (tpm) which indicates the probability of displacement from a state to another. in this study, the markov chain was used to regulate displacement probability between existing mode. the system made assumed to follow the markov chain first order. that is, to estimate mode that occurs at the present time only needs mode is going to take place in one previous unit. the process estimation parameter is done with the help of the algorithm help expectation – maximization (em) to look for the distribution parameters gaussian and tpm chain markov for each mode suryadjaja et al. modeling traffic flows with fluid flow model | 4 traffic flow. there are ten parameters that need to be estimated to compose ffm in this study, namely; • two gaussian distribution parameters for each mode, namely the average μ and the variances σ2. since the model made in this research consists of two modes, then altogether there are four gaussian distribution parameters that need to be estimated. • the π parameter indicating the probability of data flow traffic at a time is in the specified mode of. since the model made in this research consists of two modes, then altogether there are two parameters π that need to be estimated. • transition probability matrix (tpm) p which indicates the probability of migrations from a mode to other modes. since the model created on this research will classify the traffic flow into two modes, the tpm used in this study measures 2 × 2. after the required parameters are obtained from the process 0659 estimation, it can be done state estimation. estimation state is done to get the approximate value of the traffic flow based on the model compiled from the parameters that have been obtained. in this study, an estimate was made using the estimator observation and transitionbased most likely modes tracking particle filter (otpf). to test the model made, the model is applied to estimate traffic flow data on dipatiukur road, bandung city. the estimated result is compared to traffic flow data of the observation result, then calculated the value of error the estimate with method mean absolute percentage error (mape). 2.2. estimation parameters the process estimation parameter is done with the algorithm expectation – maximization (em). broadly, the em algorithm is an algorithm to suspect a parameter in a function that contains incomplete data using maximum likelihood estimation (mle). the algorithm has many advantages including this simple algorithm, based on common theories, and can be used in a wide range of applications (kusuma, a, t., and suparman., 2014), (dempster p, a., et al 2014). the em algorithm consists of two stages, namely expectation (e-step) and maximization (m-step). at the estep performed the expected expectation value for the likelihood function based on variables on the system being observed. then the mle value for the parameters is searched by maximizing the likelihood expectation resulting from estep. the parameters generated from m-step will be reused for e-step in the next iteration, and this step will be repeated until it delivers a convergent value (kusuma, a, t., and suparman., 2014). the em algorithm can generally be written as follows: here m is an index iteration, θ is a space vector parameter, y is the vector data, and fx x | is the probability density function 5 | international journal of informatics information system and computer engineering 1 (2020) 1-12 (pdf) of the complete data (moon, k, t., 1996). in this study, the equation of pdf to complete data was taken from the equation pdf distribution gaussian: ( )2, , f x ( ) 2 2 1 exp 22    − = −        x (2) where x is the observed data, μ is the average, and σ2 is a variance. by substituted the symbol on a gaussian pdf with the symbol used in this study, a pdf function generated a data flow of traffic: ( )| ,= t t f y s j 2 2 ( )1 exp 22    − − =       t j jj y (3) with yt is the number of vehicles passing in minutes tot, st is mode traffic flow in the minute to-t with jas its index θ is a space vecto r parameter that contains  2, ,j j j   . πj is the average in mode j and σj2 is a variance on mode j. the parameter πj which is the probability of data flow at a time is at mode j based on the stochastic theory markov chain. the probability of data flow traffic y is in the mode j formulated with the equation. ( , )= t p s j = j (4) by multiplying the equation (3) and (4), then summing the results for all existing state, obtained the probability equation of the data flow yt against the vector space θ: ( ), ,= t t p y s j ( | , ) ( , ) = = = t t t f y s j p s j 2 2 (y ) exp 22    − − =       t jn jj (5) ( ), t f y ( ), , = = = n t t j n p y s j 2 2 (y ) exp 22   = − − =        n t jn j n jj (6) to find the estimated value of each parameter in each iteration of the em algorithm, used lagrange multiplier. the lagrange multiplier specifies the maximum or minimum value relative of a function that is delimited by a constraint (constrain the conditions). by using lagrange multiplier, then in each iteration of the em algorithm it can calculate the estimated gaussian and tpm markovian distribution parameters at once. this can be done by making tpm constraints as a delimiter function (constraint) in the help function lagrange multiplier. ( ), ,g x y z ( ) ( ), , , , = + f x y z x y z (7) with is the auxiliary function, is the g ( x, y, z) f ( x, y, z) function that will be searched for the maximum or minimum value, and is the delimiter function  ( x, y, z) constraint). optimal value is obtained by resolving the following differential equations.   g x .0=   = =   g g y z in this research, the function that is searched for its optimal value is the likelihood function of the ( ), t f y . likelihood function which is suitable for use in the form of the log-likelihood function because the logarithmic function helps to simplify the exponential component of the ( )| ,= t t f y s j . ( )l ( )( ) 1 log , = =  t t t f y (8) where ( )l is the function log-likelihood which is searched for the minimum value there are two limitations that need to be applied to the estimated calculation of the value 0659 π. suryadjaja et al. modeling traffic flows with fluid flow model | 6 • 0 1  j any component in the tpm must be greater than or equal to zero and less than or equal to one. • 1 2 ... 1  + + + = n the entire probability value of each mode occurrence (πj) must be 1. the help function the lagrange multiplier used is: ( )j ( ) 1 2 (1 ... )    = + − − − − n l (9) the equations used to estimate the value of each parameter in each iteration of the em algorithm are derived from the decline in auxiliary function lagrange multiplier ( )j . the auxiliary function lagrange multiplier ( )j needs to be derived against the variables µj, σj 2, and π which is found in the function ( ), t f y that is searched for its optimal value. the process of deriving these equations is described as follows: ( )l     1 ( , )1 (y , ) t t t t f y f   =  =   (10) based on equations (3) and (5), then: ( ), t j f y     ( ) 2 2 1 exp 22    − − =         t j jj y ( )| , t t f y s j = = (11) while the decline of ( ), t f y  against µj and σj2 generates: ( ),    t j f y ( ) 2 2 exp 22     − − =              t jj j jj y ( ) 2 2 2 2 2 exp 2 22      − − − =           t jj t j j jj y y ( ) 2 , ,    − = =       t j t t j y p y s j ( ) 2 , , ,    − = = t j t t j y p y s j (12) ( ) 2 ,    t j f y ( ) 2 2 2 exp 22     − − =              t jj j jj y ( ) 2 2 2 2 exp 22     − − =                 t jj j j j y ( ) 2 2 2 exp 22     − − +           t jj j jj y ( ) 2 3 2 1 1 exp 2 22    − − =  −              t jj j j y ( ) 2 2 1 exp 22    − − +           t jj j j y ( ) ( ) 2 4 2   − −  − t j j y ( ) 2 2 exp 2 2      − − =               t j j j j y ( ) 2 2 4 1 2 2    −  − +         t j j j y ( ) ( ) 2 2 4 1 , , . 2 2     − − = − + =         t j j t t j y p y s j (13) so, the equation (10) can be written as: ( )    j l ( ) 1 1 | , , (y , )  = = = t t t t t f y s j f (14) ( )    j l ( ) ( ) 2 1 1 , , , ,    = − = = t t j t t t t j y p y s j f y (15) 2 ( )    j l ( ) ( ) 2 2 4 1 1 1 , 2 2     − = − = − +          t t j j t t j y f y ( ), , . = t t p y s j (16) according to the theorem of bayes depicting conditional opportunities 7 | international journal of informatics information system and computer engineering 1 (2020) 1-12 between the two occurrences of a and b are as follows: ( )|p a b ( ) ( ) ( ) | ,= p b a p a p b (17) then it can be done forward filtering ie counting chances of mode jbased on traffic flow data yt based on the following equation: ( | , )= t t p s j y ( ) ( ) , , ,   = = t t t p y s j f y ( ) ( ) | , . ,    = = j t t t f y s j f y (18) next, backward filtering is performed, namely smoothing process on the probability function that is generated from the forward filtering process. calculated value 1 ( | , ) + = t t p s j y until the value of 1 ( | , ) + = t t p s j y is obtained. the equation is used as follows [8]: 1 ( | , ) + = t t p s j y ( ) ( ) 2 1 | , = = = m kj t t j p p s j y (19) then the following computed probability function backward filtering: for 1, 2, ,1= − −n n n count [8]: ( | , )= t t p s j y ( ) ( ) 2 1 1 | , | ,  = + = = =     t t j n t p s j y p s k y ( ) ( )1 | ,+ = m t t jk p s k y p (20) based on the equation (18), then the equation (14), (15), and (16) can be written to: ( )    j l ( ) 1 1 | ,  − = =  = t j t t t p s j y (21) ( )    j l ( ) 2 1 | ,   = − = = t t j t t t j y p s j y (22) 2 ( )    j l ( ) 2 2 4 1 1 2 2    − = − = − +          t t j j t j y (23) ( )| , = t t p s j y the optimal value of the equation lagrange multiplier can be found by creating the first instance of the auxiliary function against each delimiter function variable equal to zero. as a result, equations (21), (22) and (23) need to be made equal to zero. by making the equation (22) equal to zero, then: 1 ( | , ) =  = t t t t t y p s j y ( ) 1 | ,  = = = t j t t t p s j y j  1 1 ˆ( | , ) . ˆ( | , )   = = = = =   t t t t t t t t t y p s j y p s j y (24) by making the equation (23) equal to zero, then: ( ) 2 1 ˆ ( | , )  = −  = t t j t t t y p s j y ( ) 2 1 | ,  = = = t j t t t p s j y 2 j  ( ) 2 1 1 ˆˆ ( | , ) ˆ( | , )    = = − = = =   t t j t t t t t t t y p s j y p s j y (25) based on equations (9) and (21), then: ( )    j j ( ) 1 1 ,| , 0   − = = = − = t j t t t p s j y or can also be written as: ( ) 1 | , = = t t t t p s j y .= j (26) suryadjaja et al. modeling traffic flows with fluid flow model | 8 with a sum of equations (26) as much as 1, 2, , ,=j n generates: ( ) ( )  1 | , ... | ,  = = + + = t t t t t t p s j y p s n y   1 ...  = + + n or: 1 :j = ( )  1 1 1 | ,  = = = t t t t p s y 2 :j = ( )  2 1 2 | ,  = = = t t t t p s y … :j n= ( )  1 | ,  = = = t t t n t p s n y + ( ) 1 1 | , = = = t n t t t j p s j y 1 2 ( ... )   = + + + n because ( ) 1 | , = = n t t j p s j y and 1 2 ...  + + + n are worth 1, then:   1 1 =  t t ( )1= 1 1 ... 1+ + + t ( )1= 1t  ( )1= t = by substituted λ with t on the equation (26), it generates the equation:  j ( )1 1 ˆ| , − = = = t t t t t p s j y (27) transition probability matrix (tpm) p created in this study was a 2 × 2-sized matrix, indicating the probability of switching between mode. to look for the tpm value, it takes the probability value generated from the backward filtering process 1 ( | , ) + = t t p s j y and ( | , )= t t p s j y for each of the entire data generated from the observation process, except for the data from the first minute. it also takes tpm values from previous iterations of the em algorithm. with this information, in each iteration of the em algorithm it is possible to estimate the value of the tpm element in row j and column k with the following equation taken from reference (sutarto, y, h., and joelianto, e., 2015) with minor modifications after discussion with the reference author: ( )m jk p ( ) ( ) ( ) ( ) ( ) 1 1 2 1 1 2 | , | , | , . | ,     − − = − − = = = = = =   m t jk t t t t t t t t t t t p p s j y p s k y p s k y p s j y (28) the process estimation parameter with the em algorithm that has been described is done with the help of the computer. at each iteration, the first calculation of the function log likelihood ( )l  by equation (8). specifically, for the first iteration, used the parameters µj, σj2, and πj. which is initial condition. initial condition to µj and σj2 is determined from the average and the variance of data from the observation result of the first observation point for each mode. initial condition for πj is considered balanced for both modes, so πj respectively, that the respective mode is 0.5. initial condition for the matrix p is considered balanced for the whole possible displacement of mode, so the initial condition for the entire matrix element p is worth 0.5. the initial condition that is used in the process of the estimation parameter is shown in table 1. performed forward filtering with equations (18) and backward filtering with equations (20). then, it is done with 9 | international journal of informatics information system and computer engineering 1 (2020) 1-12 the calculation to find the values µj, σj2, and πj, new using the equation (24), (25), and (27). also calculated tpm values with equations (28). then the value is recalculated and calculated the difference with the value ( )l  from the previous iteration. the iteration process is repeated continuously until the difference in value ( )l  with the value ( )l  (threshold). this study used the threshold of 0.00001 due to multiple attempts, after a difference in value of ( )l  is no larger than the threshold value of the estimated parameters no longer experiencing a change in the value significant. 2.3. state estimation after getting the parameters of each mode and making the initial estimate without estimator based on those parameters, the next step that needs to be done is the state estimation process (state estimation). state estimation is the process of looking for the best aproxation regarding state of a system based on known information. in this study, state estimation was used to get estimated traffic flows based on the value of parameters that have been generated by the em algorithm. the process state estimation is generally done with the help of estimator. the estimator of used in this study is a sequence of particle filter observation and transition-based most likely modes tracking particle filter (otpf). otpf is a particle filter variant that is suitable for a system consisting of several modes. otpf considers mode to as an unknown parameter. therefore, on otpf there is a stage that does not exist in the usual particle filter, which is the stage to do the estimation of mode that is most likely to occur from any particle. once the approximate of is being obtained is most likely to occur, the system is assumed to follow the dynamics of the corresponding mode. the next stages of particle filter are done as usual (kusuma, a, t., and suparman., 2014). mathematically, the otpf algorithm can be written as follows (kusuma, a, t., and suparman., 2014): 1) initialization • 0t = . the mode at time 0 is describe as m0. for 1, ,i n= , the ustron (1) 0 x from the initial condition (initial condition) in the 0 m mode then set 1t = . 2) prediction • for each mode j t m which has a displacement probability 1− j t t m m t from mode 1t m − to j t m that value is not equal to zero ( 1, ,j k= with k represent the number of modes), sampling ( )( ) (1)1| −j t i t t tm x p x x , for 1, ,i n= . • for each mode j t m , it is calculated importance weight ( )( ) (1)|j t i t t tm w p y x , for 1, ,=i n . 3) mode selection • calculated the average particle weight in each mode j t m then multiply by the probability of mode switching: 1 (i) 1 / − = =  j t j t t n m t tm m i w t w n . • find the most likely mode: arg max := j t j t m t tm m w for all , 1, ,=jtm j k . suryadjaja et al. modeling traffic flows with fluid flow model | 10 • normalize the weight of particle in mode j t m . 4) resampling • resample n new particle  ( ) ,itx to 1, ,=i n replace the particle in mode  ( ) ,it tm x to 1, ,=i n base on importance weight. • set 1= +t t then return to stage 2. 3. results and discussion this section explains the testing of the accuracy of the model by applying the model to estimate the actual traffic flow data on dipatiukur road, bandung. to find out the accuracy of the resulting estimation, the value of error estimation is calculated using the mape method 3.1. retrieval of data at this stage traffic flow data is collected on the observed road section. the data collection process was carried out on monday 20 september 2017 at 06.0010.00. data is collected by counting the number of vehicles that pass each minute manually with the help of the application counter contained on android smartphone. observation of the traffic flow is carried out for both directions of traffic on the observed road, i.e. the flow from south to north and flow from north to south. observations are made without distinguishing the types of vehicles that pass, so that all vehicles that pass the observation point are included in the calculation. observations were made at two ends of the road that were observed simultaneously. the traffic flow data per minute is collected into traffic flow data per three minutes. by selecting a longer data interval, it can be minimized the occurrence of random noise in the data so as to reduce the value of error resulting in the estimation process (yao-jan., et al, 2012), (moon, k, t., 1996). the first observation point is in front of the ithb campus while the second observation point is in front of indomaret point dipatiukur. location of the observation of traffic flow can be seen in fig. 1. fig. 1. location of traffic flow observations 3.2. model testing after the traffic flow data on the observed road sections have been successfully collected, it can be tested on the modeling methods designed in this study. the em algorithm is applied to estimate the parameters of traffic flow data to the north and south. the parameters generated are shown in table 1. after the required parameters have been obtained from the parameter estimation process, then we can estimate the traffic flow using otpf with 500 particles. the estimation results are shown in figs. 2 and 3. visually it can be seen that the resulting estimation results are quite close to the traffic flow data from observations. in this study, to determine the numerical accuracy of estimation 11 | international journal of informatics information system and computer engineering 1 (2020) 1-12 values, the calculation of the error value estimate using the method mean absolute percentage error (mape). in general, the amount of mape can be used to classify the accuracy of a forecast into four levels, namely (kumar, s, v., 2017): • if mape is less than 10%, the accuracy is highly accurate • if mape is between 11% and 20%, the accuracy is good. • if mape is between 21% to 50%, the accuracy is reasonable. • if mape is more than 51%, the accuracy is included in inaccurate category. the estimated traffic flow generated in this study has a mape value of 7.7978% for traffic flow from south to north and 5.8547% for traffic flow from north to south. both of the mape values are below 10%, so the estimation accuracy generated by the model belongs to the category highly accurate. table 1. parameter values generated from the em algorithm. paramete r traffic flow to the north traffic flow to the south mode 1 mode 2 mode 1 mode 2 µj 83,4794 90,3488 76,0913 78,102 3 σj2 183,987 7 1208,944 7 171,961 2 17,091 2 πj 0,7040 0,2960 0,7719 0,2281 pjk 0,9995 0,0005 0,9897 0,0103       0,9893 0,0107 1,0000 0,0000       fig. 2. comparison of the results of the estimation experiment using otpf with the observed data from south to north traffic flow fig. 3. comparison of the results of the estimation experiment using otpf with the observed data from north to south traffic flow 4. conclusion based on parameter estimation, estimation state and system testing that has been carried out in this research, it can be concluded that the model compiled from mixed gaussian distribution with parameters generated by parameter estimation with em algorithm can already accurately describe the traffic flow. with an estimated mape value of 7.7978% and 5.8547%, it can be said that the estimate has a high accuracy, and includes the category highly accurate. the mape value is even lower than the traffic flow model produced by suryadjaja et al. modeling traffic flows with fluid flow model | 12 other studies that are more complex and require more data. there are still a few gaps that allow even more accurate modeling. some of them are by increasing the number of modes, extending data intervals, and increasing the amount of data used in the model testing process. models can also be drawn that describe dynamic systems such as the autoregressive (ar) model to get estimates with higher accuracy. references a.s. wiyono. (2014, okt 22). bandung kota termacet ketujuh, ini penyebabnya [online]. available: https://www.merdeka.com/peristiwa/bandung-kotatermacet-ketujuh-ini-penyebabnya.html ali, m. (2011, april). intelligent transport system (its). in 10 th research seminar series workshop. clark, s. (2003). traffic prediction using multivariate nonparametric regression. journal of transportation engineering, 129(2), 161-168. dempster, a. p., laird, n. m., & rubin, d. b. (1977). maximum likelihood from incomplete data via the em algorithm. journal of the royal statistical society: series b (methodological), 39(1), 1-22. kumar, s. v. (2017). traffic flow prediction using kalman filtering technique. procedia engineering, 187, 582-587. kusuma, t, a., & suparman. (2014) algoritma expectation-maximization (em) untuk estimasi distribusi mixture. jurnal konvergensi, 4(2), 65-83. marinică, n., & boel, r. (2012, june). platoon based model for urban traffic control. in 2012 american control conference (acc), 6563-6568. moon, t. k. (1996). the expectation-maximization algorithm. ieee signal processing magazine, 13(6), 47-60. salim, t., hutagalung, m., & sutarto, h. y. (2019). simulasi arus lalu lintas menggunakan perangkat lunak simulink. jurnal telematika, 13(1), 7-12. sutarto, h. (2016). adaptive control for traffic signals using a stochastic hybrid system model (doctoral dissertation, ghent university). sutarto, h., & joelianto, e. (2015). expectation-maximization based parameter identification for hmm of urban traffic flow. international journal of applied mathematics & statistics, 53(2), 90-101. tafazoli, s., & sun, x. (2006). hybrid system state tracking and fault detection using particle filters. ieee transactions on control systems technology, 14(6), 10781087. vandaele, n., van woensel, t., & verbruggen, a. (2000). a queueing based traffic flow model. transportation research part d: transport and environment, 5(2), 121-135. wu, y. j., chen, f., lu, c., smith, b., & chen, y. (2012). traffic flow prediction for urban network using spatio-temporal random effects model. in 91st annual meeting of the transportation research board (trb). https://www.merdeka.com/peristiwa/bandung-kota-termacet-ketujuh-ini-penyebabnya.html https://www.merdeka.com/peristiwa/bandung-kota-termacet-ketujuh-ini-penyebabnya.html 25 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 sustainability policy in indonesia: case study economic structure and determinants in banjar municipality agus supriyadi1, tao wang2, melda ria juwita3, rayhan gunaningrat4, susi safitri5, giuseppe t cirella6 1,2,3,5the school of geography, nanjing normal university, nanjing 210023, china 4program studi manajemen, fakultas hukum dan bisnis, universitas duta bangsa, indonesia 6faculty of economics, university of gdansk, 81-824 sopot, poland correspondence: e-mail: agus_dea@yahoo.com a b s t r a c t s a r t i c l e i n fo in response to governance shift in indonesia, a heavy burden of responsibility is placed on local authorities to act in the best interest of its citizenry. this study examined evidence of local government policy, economic structure, and economic growth in the new autonomous city of banjar municipality, west java province, indonesia and assessed determinants of regional transformation. the methodology used location quotient, analysis, shift-share analysis, and ordinary least square regression analysis. the findings revealed in the ten-year period of the study on agriculture, business trade, accommodation, other sectors and public service dominated the structure of the economy. in terms of the authority, the municipality’s policy of sustainability is aimed at protecting the traditional market and developing local road infrastructure—contributing to its economic growth. sustainability-oriented policies are designed to stress order, control, and accelerate economic activities which regional authorities sometimes considering them as challenging due to transitional requirements and demands. article history: ___________________ keywords: sustainable development, economic growth, autonomy city, gevernance, indonesia. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 2(1) (2021) 25 46 received 10 may 2021 revised 20 may 2021 accepted 25 may 2021 available online 26 june 2021 agus supriyadi, et al. sustainability policy in indonesia: case study … | 26 i. introduction indonesia is one of the most populous and largest archipelagic countries in the world, with multi-ethnic, multi-cultural, and various regional languages facing a number of management-oriented challenges. since the economic crisis of 1997 hampered most of asia, many countries have taken action in response to the condition, including indonesia. the crisis led the indonesian government to implement transformative centralized to decentralization action, where some of the state powers and resources were transferred to the lower level of government, i.e., at the provincialand municipal-level. in 1999, a policy known as regional autonomy was implemented, converting the country into the largest worldwide laboratory, to test out the decentralization experiment (miller, 2013). as a response, the success of this policy move on economic and income growth still depends on the local government’s response to the market and the orientation of economic development asia-wide (asian development bank, 2010). a significant issue that usually arises in a number of countries is the dependency and burden of local governments highly reliant on central government transfer (i.e., by not maximizing local resources and revenue) (masaki, 2019; kelly, 2012; oates, 1993). evidence from ghana in 2004 has shown that higher external transfers toward local authority does not encourage the increase of local revenue. this condition contradicts one of the fundamental motivations of the decentralization policy, where the effectiveness and efficiency of allocated local resources (i.e., by the local authority), can be outweighed by local knowhow (e.g., indigenous knowledge and community needs) (mogues & benin, 2012). the purpose for regional development is a bottom-up approach to improving the communityin terms of fiscal decentralization by enhancing economic efficiency and social wellbeing (oates, 1993). a useful example from china illustrates cities with higher gross regional domestic product (grdp) tend to be more inclusive, i.e., cities with a larger population and higher grdp per capita have a significant industrial sector contributing to the gap of inclusiveness between the high and low skilled (hu & wang, 2019). prior research has also shown that in terms of economic growth, human capital contributes significantly compared to fixed capital investment (liu et al., 2018). as a result of indonesia’s reformation, regional autonomy was followed up by a number of responsibilities in which local governments highly accommodating their new-found power. the system, however, brought forth both opportunities and challenges (miller, 2013; sondakh & jones, 2003). the transferability of authority, to date, directly associated local governments with the underlying objectives of poverty reduction, public service improvement, and public participation— something the central government has also deemed as critical (suryahadi et al., 2012). banjar municipality an autonomous city located in west java province with a population of more than 200,000 inhabitants. the city was allocated as one of the new autonomous cities established at the implementation of the regional autonomy policy— nationwide. since its inauguration in 2002, there has been massive development in terms of infrastructure and public service. in 2010, based on an assessment of 33 indicators, with four main factors of assessment (i.e., community welfare, governance, public services, and regional competitiveness), 27 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 banjar municipality has achieved awardlike success as the second most thriving autonomous city in west java province. despite the location of the city, at the east border of west java province and faraway from a metropolitan city, its performance in accordance with good governance has assisted in piecing together a highly productive government output. positive economic growth of the city, however, has never reached more than 5.6% (bps-statistic of banjar municipality, 2017) which has some cause for concern—especially in terms of policymaking. since its establishment, research into its leading sectors as well as economic structure, i.e., concerning economic growth, and how the municipality by protecting the traditional market has not been done. evidence of previous research from the state of michigan—and municipalities within it—shows incentive-based policies that play an important role in employment and in establishing new businesses. these policies have been considered one of the most successful strategies in the postcrisis era (drucker et al., 2019). based on this economic development, we consider local regulation and analyze independent variables to best calculate and explain economic growth in banjar municipality (see fig. 1). banjar municipality’s uniqueness, as a city, was inaugurated by the minister of the home affairs as one a municipality that administratively controls not only urban but also rural villages. this important fact internalized our research to focus on providing important guidance and recommendations in this context. urban villages (kelurahan) are territorial units occupied by a number of residents, at the lowest governmental level, are organized directly under the district head and have no rights to household ownership. the main characteristic of the urban village is that it is led by civil servants and not chosen by the people. local authorities can postpone a permit, for example, when investigating developmental ideas and projects. as such, a modernized market was introduced to help with a lagging traditional market during the economic crisis—noted by trade sectors indicators and revived vital points in economic development. the regional autonomy policy in indonesia successfully spurred research interest in trend, impact, determinant, and impact on economic growth; however, limited studies have looked at how these new autonomous cities operate independently. the paper conducts critical research into better understanding the problems and challenges of the state of the art of banjar municipality and presents future solutions. we fill in a number of gaps, in terms of sectoral information and economic structure by relating our findings to economic growth in the municipality. the scope of the research will benefit policymakers and city planners alike as well as scholars in helping guide and benchmark future works. a breakdown of the paper is structured as follows: section 2 is a literature review, section 3 contains the methods, section 4 illustrates the results, section 5 elucidates a discussion on impact and development, and section 6 presents the conclusions and policy implications. agus supriyadi, et al. sustainability policy in indonesia: case study … | 28 fig. 1. economic growth of banjar municipality compared with regional and national levels, 2007-2016. 2. theoretical framework 2.1 globalization and regional autonomy the restructuring of indonesia’s economy is the primary factor for its economic growth (lankauskienė & tvaronavičienė, 2013). lessons from china and india play an important role in the region in terms of labor and influence (mallick, 2017). this mobility assert transformative effects during the period between 1970 to 1990; as such, structural changes on average had not been conducive to productivity growth (fegerberg, 2000). globalization and the development of technology in the twenty-first century significantly 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 banjar municipality west java province cimahi municipality national 29 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 changed the global landscape and palpably contributed to its transformation. a focus on economic development is now transforming job growth and businessto job creation, business growth, and income growth (i.e., factors that are the typical and reasonable within indonesia’s economic development policy). nonetheless, broader social and environmental improvements parallel the strengthened economy (moore et al., 2006). regional autonomy came into play, around the world, as a possible solution to solving regional economic problems. decentralization theory, as the foundation of regional autonomy policies, leans toward a robustness in comparison to a centralized country (paganelli, 2006). compelling empirical evidence from indonesia, since regional autonomy was implemented to 2015, showed regional revenue (i.e., nationwide) increased by 374%, a significant impact and confidante with the policy (suryani, 2015). key factors of economic growth are discussed by wagner’s theory of “law of expanding state expenditure” in which increasing per-person income in correlation with increases in state expenditure (sobari, 2011; wagner, 1890; suryani, 2015) can prove to be beneficial. research conducted using location quotient (lq) and shift share analysis (ssa) in banyuwangi regency in east java province revealed that one overpowering sector does not necessarily give overarching value to a regional economy (khusaini, 2015) due to a lack of economic balance and sustainability—cutting local competitiveness and creating monopolylike circumstances (akib et al., 2019). as such, it is essential to consider the quality of human capital, including: wellbeing, distance and travel time to work, transportation service, road and infrastructure condition, and other supporting amenities (endaryanto et al., 2015). 2.2. regional economics and gross regional domestic product (grdp) dunn discussed a statistic approach to understanding regional economics and its contributive effects via differing sectors’ contribution to economic growth (dunn, 1960). his research continues to influence researchers in various fields, including a number of studies, that used ssa to explain certain economic phenomena in a community (e.g., regional employment development trend and regional growth (goschin, 2014; rigby & anderson, 1993). other examples of ssa include a spanish study that revealed higher competitiveness in agriculture and construction sectors, while industry and service had the opposite effect (fernández & menéndez, 2005). three components that are widely used include: (1) national and regional share components, (2) industrial mix components and proportional shift, and (3) differential and competitive share components (evans, 2008; noisirifan, 2018; morrisey, 2014). in a new paradigm of regional economic theory there is a basic developmental shift (i.e., transformation) from development economic sectors to economic institutions (kifli & umami, 2017). in the globalization era, autonomous cities makeup one form of the implemented economic institutions’ framing. despite its critics, the use of gross domestic product (gdp) as a measurement of human wellbeing has held fast for half a century—accepted as the standard measure for a country’s progress (contanza et al., 2009). in the local economy, grdp generally uses agus supriyadi, et al. sustainability policy in indonesia: case study … | 30 indicators that measure macroeconomic activity—locally (andayani, 2018; feriyanto, 2014). several studies, however, used grdp and revealed how leading local products and spatial economic contributions (i.e., regionally and nationally) showed meaningful and reciprocal economic results (kristiawan et al.,2016; kifli & umami, 2017) ; endaryanto et al., 2015). johnston and huggins also highlighted that a variety of industrial structures can significantly affect the dynamism of regional growth (johnston & huggins, 2018). the implementation of regional autonomy in indonesia signaled new authority to control and develop their local needs. prior research from china showed how low quality of rural roads can significantly influence economic growth (i.e., in a negative matter compared towell-developed ones) (fan & chan-kang, 2005). hence, built infrastructure policy is one important way of maintaining positive economic development. in terms of sustainable economic growth, investment in education is also one key policy that should be considered for local governments (liao et al., 2019; sukamdi, 2014). evidence from the 1970s and 1980s show a positive correlation between government expenditure and economic growth—investment playing a large part of the growth (bose et al., 2007). another variable that contributes significantly to the expansion of an economy is its labor force, as can be seen in the united states from 1948 to 2001 with an increase of 1.7 % per annum to the average annual growth of potential real gdp (daly & regev, 2007). this type of advancement also occurred in china between 1990 and 2009 where its economy stabilized its growth by promoting and improving labor quality and consumption capabilities (zhu et al., 2011). reinforcing this viewpoint, nosier and el-karamani stress a series of variables such as education, health, government consumption, and trade openness as having a nonmonotonic effect on growth through education and physical capital accumulation via labor (nosier & el-karamani, 2018). moreover, resolving labor force and unemployment problems for urban and rural areas will require different parametric expertise (i.e., urban policies should focus on service sectors while rural should be on agricultural ones) (suryadarma et al., 2013). it is also essential for a country to invest in human capital (ali et al., 2018) to stimulate economic activity for best policymaking practices—developed via appropriate regulatory and local knowhow (fisher, 1997; romyen et al., 2019). from this outlook, our paper’s scope prepares sectoral-based research that measures banjar municipality’s transition as an autonomous city and bestows future economic requirements and demands for the betterment of indonesia and its policy development. 3. data source and methods 3.1. data source this research examines leading sectors, economic structure, and grdp in the new autonomous city of banjar municipality, indonesia. it is located on the east border of west java province. the municipality’s population has grown from 160,810 in 2002 to 202, 362 in 2017 (bps-statistic of banjar municipality, 2017). the city is expected to grow in terms of industrial development, trade, service, and tourism via the east region of west java. since its inauguration, the municipality has had to uniquely configure its authority over its four districts (i.e., kecamatan) as well as its 22 villages and urban villages (i.e., desa and 31 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 kelurahan). this makeup has fashioned banjar municipality as one of the new autonomous cities administratively as a city and municipality that also manages village and urban village areas. since it was established the municipality’s revenue has been dominated by trade, hotels, and restaurants. the method of the data collected for this study utilized two methods. first, we collected online data from bps-statistics banjar municipality’s website and, second, we directly requested the data from the local authorities, including: regional planning and development agency of banjar municipality and regional financial and asset agency. the data period spans from 2007 to 2016. we utilize grdp sector data from bpsstatistics banjar municipality which is grouped and categorized based on the system of national accounts 1993 and indonesian standard industrial classification (i.e., klasifikasi lapangan usaha indonesia) 1990 (bps, 2017). the land use map of banjar municipality in 2006 acts as a starting point for this research (see fig. 2). as can be seen in figure 2 agricultural areas dominate almost in many areas. the center of urban area growth is still dominated in the city center, but the agglomeration of urban areas is divided into two regions namely in the banjar district and the langensari district. this is in line with the concept planned by the local authority that expects two centers of economic activity to grow. we further adjust via the international standard industrial classification for all economic activities which were released by united nations sources. the grdp sectors include: agriculture, mining and quarry, construction and building, processing industry, electricity, gas and water, transportation and communication, trade, hotels and restaurants, and service sectors—all considered key economic activities (arent et al., 2015). variable symbols, descriptions, and data sources are shown in table 1. agus supriyadi, et al. sustainability policy in indonesia: case study … | 32 fig. 2. land use map of banjar municipality (supriyadi et al., 2020)) 33 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 table 1. variable and data source variable name variable variable description data source agriculture, livestock, forestry, and fishery agr includes farm food crops, non-food crops, livestock and product, forestry, and fishery bps-statistics banjar municipality mining and quarrying mc includes non-oil gas and quarrying bps-statistics banjar municipality manufacture industry mi non-oil and gas manufacturing bps-statistics banjar municipality electricity, gas, water supply, and water treatment egw electricity, gas water supply, and treatment consumptions bps-statistics banjar municipality construction and building cons construction activity bps-statistics banjar municipality trade, accommodation, and food business taf wholesale and retail trade, hotels, and restaurants bps-statistics banjar municipality transportation and communication tc transportation road, railway, air, and communication bps-statistics banjar municipality finance and business services fbs includes finance, rent, and businessservices bps-statistics banjar municipality other services os public and private services bps-statistics banjar municipality 4. research methods there are plenty of methods that can describe a leading sector and the structure of the economy of a city. however, this paper primarily utilized lq and ssa. in addition, to explore the main factors from the decentralization sectors that affected economic growth we utilized ordinary least square (ols) regression analysis. 4.1 leading sector analysis to get a specific description of industrial concentration in an area (i.e., by decomposing where a product is in a particular geographical place) we comparing them to wider areas using lq analysis which is commonly used as a quantitative technique in base economic theory (moore et al., 2006). in terms of the changeability of the leading sector’s calculation, this paper uses a formula, i.e., eq. (1), used by several previous studies (morrisey, 2014; crawley et al., 2013; riddington et al., 2006; jahn, 2017). the starting point collects the distribution of grdp at the constant market price from 2007 to 2016 and calculates it accordingly. where: lq is the value of the leading sector in one geographic area (note: lq > 1 is defined as the leading sector; lq < 0 is defined as a sector in deficit and the area might be relying on this sector from a different area; lq = 0 is defined as a sector only supply for the local area with no potential export); 𝑔𝑖 𝑡 is the value of sector i in one geographic area in the year 𝑡; 𝑔𝑡 is the sum of all the sectors in the agus supriyadi, et al. sustainability policy in indonesia: case study … | 34 year 𝑡; 𝐺𝑖 𝑡 is the value of the sector i in the regional geographic area in the year 𝑡; 𝐺𝑡 is value of the whole sector in the regional geographic area in the year 𝑡. the ssa is one method that is widely used and credited as a tool to understand political economy, retail analysis, and regional economy (knudsen, 2000). a ssa is also applicable in its analysis to measure the implication of structural development in a country’s international competitiveness—over time—when it was applied to export growth (wilson et al., 2005). despite the weakness of the method, since it does not explain economic growth patterns, ssa has been proven popular for almost half a century due to its very straightforward logic and interpretation (jackson & haynes, 2009; dogru & sirakaya-turk, 2017). three components makeup the ssa (i.e., regional share (rs), industrial mix (im), and competitive share (cs)) formulated in eq. (2). where: 𝑆𝑖 is the value of ssa for sector i, 𝑅𝑆𝑖 is regional share represented by eq. (3); 𝐼𝑀𝑖 is a mix component in sector i represented eq. (4); 𝐶𝑆𝑖 is a competitive component in sector i represented eq. (5). where: 𝑔𝑖 is the value of sector; t is the year of last period of the research; n is the sum of the research period. 4.2 regression analysis based on the reviewed literature, this study develops a linear regression model using ols regression analysisi to determine variables that have a significant contribution to the economic growth of banjar municipality (kowalski, 2000; greene, 2012; glodowska & pera, 2019). the variables used are components that the local government has authority over, that being: local government expenditure, local regulation, labor force, education facilities, and local road facilitation. these independent variables once pieced together formulate eq. (6). where: eg is economic growth as a dependent variable; 𝛼0 is the intercept parameter; 𝛽1,..𝛽𝑘 are regression parameters; 𝐸𝑥 is local government expenditure; 𝐿𝑅𝑒𝑔 is the local regulation; 𝐸𝐹 is primary and secondary education facilities; 𝐿𝑅 is the local road; 𝜀 is the error term. the significance test conducted also determined the value of the independent variable. hence, r2 and adjusted r2 values that are closer to one are good indicators of high correlation and a pvalue indicating the degree of confidence is set to 5 % or less. next, a multicollinearity test was performed to determine if there was a correlation between the independent variables. if a correlation was present, then a variance inflation factor (vif) of greater than ten would indicate a positive multicollinearity and the independent variable would yield affective. 5. results the economic growth of banjar municipality was relatively stable during 35 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 the research period, with minor positive and negative fluctuation pending the variable. an important lesson in retrospect of china’s economic structure reminds us of similar results in that country’s early years before its success as an economic force (peng et al., 2017). in the first five years of banjar municipality’s anonymity, the city’s policy focused on infrastructure development (e.g., local government offices, public service facilities, and relating amenities). these initial investments proved useful and lay the groundwork for its potential growth. 5.1 potential growth from economic sectors when banjar municipality was allocated as an autonomous region it had limited infrastructure, public facilities, and out-andout authority (i.e., political clout) on its population. as such, it is important to recognize crucial urban planning requisites that consider availability of resources and ability to compete with other regions. banjar municipality does not have natural resources and, thus, its revenue comes from other economic sectors. in this study, lq and ssa were used in terms of grdp, broken down into sectors, from bps-statistics banjar municipality data (bps-statistic of banjar municipality, 2017). the lq result is illustrated in table 2 in which three main potential economic sectors contribute to the development of the city, namely: (1) agriculture, livestock, forestry, and fishery; (2) trade, accommodation, and food business; and (3) finance and business services. notably, other services also have a significant impact. based on the result, agriculture, livestock, forestry, and fishery has the greatest potential for export. in terms of location, the municipality is traversed by the citanduy river which could be one of the specific reasons for a strong agricultural sector— acting like the backbone of the economy. trade, accommodation, and food business as well as finance and business services also performed well as base sectors. inference on their success, again, pinpoints on the municipality’s location in which the intersection national road to central java province and the tourism area of pangandaran beach play an important role. the other services, noted, consist of public and private sector advancements such as health services that currently dominate the economic growth with the founding of the city’s a-type hospital facility (i.e., one of the most modern hospitals in the region). in general, construction and building as well as transportation and communication were two sectors that experienced a decreasing trend in the tenyear study. in terms of ssa, each economic sector was analyzed by calculating the grdp of each sector using the overall regional sector score. data originated from banjar municipality and west java province over a period of ten years from 2007 to 2016. the result of the ssa analysis can be seen in table 3. agus supriyadi, et al. sustainability policy in indonesia: case study … | 36 table 2. lq values for each economics sector lq variable 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 agr 1.57 1.59 1.45 1.51 1.59 1.57 1.56 1.54 1.59 1.55 mc 0.13 0.13 0.12 0.12 0.13 0.11 0.12 0.12 0.12 0.12 mi 0.26 0.25 0.27 0.28 0.25 0.26 0.25 0.25 0.24 0.24 cons 1.54 1.54 1.56 1.47 1.18 1.17 1.19 1.19 1.16 1.15 egw 0.48 0.49 0.45 0.45 0.80 0.45 0.43 0.43 0.46 0.46 taf 1.62 1.67 1.60 1.94 1.54 1.63 1.66 1.66 1.64 1.64 tc 1.65 1.72 1.62 1.45 1.88 1.15 1.16 1.17 1.18 1.17 fbs 2.16 2.11 2.02 1.92 1.94 1.79 1.76 1.75 1.72 1.67 os 2.16 2.27 2.22 2.13 2.09 2.06 2.06 2.16 2.16 2.17 table 3. comparative advantage of ssa values for each economic sector between 2007 and 2016 sector lq† rs im cs ssa total agr yes 144,960.41 1,717.99 111,798.69 31,443.73 mc no 3,256.11 5,444.92 2,463.41 274.59 mi no 119,449.19 6,483.90 74,696.55 38,268.74 cons no 3,012.98 486.20 2,492.06 1,007.54 egw no 99,828.02 12,642.33 58,590.73 53,879.62 taf yes 330,510.67 28,907.31 155,001.84 146,601.52 tc no 88,763.09 75,531.91 66,229.15 98,065.85 fbs yes 73,581.70 17,993.61 49,218.70 42,356.62 os yes 182,678.69 45,039.53 112,341.86 115,376.35 according to table 3, it is can be said that all of the ssa totals retained a positive value, meaning those sectors developed a positive, contributing and absorbing, labor force. the highest contributing variable to economic growth and employment was trade, accommodation, and food business with †supporting sector of economic development; source: authors own elaboration 37 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 330,510.67 of the total ssa value and the lowest was mining and quarrying with 274.59. agriculture, livestock, forestry, and fishery; trade, accommodation, and food business; and other services have a significant positive value in the regional share equating to these contributing sectors as lead absorbers of employment and economic growth. these sectors, considered as first level sectors, that should be maintained. in the industrial mix, it can be highlighted that construction and building; electricity, gas, water supply, and water treatment; finance and business services; and other services have a general positive value. it is interesting to note, with the exception of mining and quarrying, none of the other sectors had a positive competitiveness when combined with a provincial level supply chain, meaning there were no noticeable specialties found municipality-wide and no other autonomous entities (e.g., regions or cities) that could be can relied upon for economic growth. 5.2 regional autonomy and economic growth variables based on the literature, this study examined influences that significantly contributed to the economic growth of banjar municipality. we applied five factors to model the ols regression analysis. the factors were chosen based upon the local government authority’s implementation of regional autonomy and its relating policy adjustments. the ols regression analysis results are illustrated in table 4. the model was analyzed using stata 15 software to check for correlative and determinative variable results and show if they are representative of reality. the result of the model showed that the multiple adjusted r-squared value was 0.8070 which is nearer to 1.0—indicating an overall strong relationship. the process then checked the multicollinearity of every independent variable (i.e., a vif < 10 meant there is no multi-collinearity in the regression model). based on the result each factor had value < 10, hence, concluding that the regression model had no multicollinearity for all of the independent variables. 6. discussion for the ten-year study from 2007 to 2016, it can be deduced that the sectors that contributed to banjar municipality’s economic development were agriculture, livestock, forestry, and fishery; trade, accommodation, and food business; finance and business services; and other services—in line with previous research conducted in aceh province (kifli and umami, 2017). the location of banjar municipality, the suburbs of west java province, yields this city closer to rural areas and, hence, subject to agricultural products and related outputs. generally, the agriculture, livestock, forestry, and fishery sector have an essential role in economic growth; however, compared to other regions, the impact of the agricultural output is relatively insignificant since the city has a small geographical area. agus supriyadi, et al. sustainability policy in indonesia: case study … | 38 table 4. ols regression analysis independent variable standard error significance (p value > t) † r-squared adjusted rsquared vif expenditure 0.057035 0.090 ** 5.68 local regulation 0.51662 0.034 ** 0.9142 0.8070 3.28 labor force 0.470402 0.693 5.96 education facility 1.571344 0.083 ** 9.07 local road 7.055934 0.040 ** 9.48 historically, before banjar became an autonomous region, it was a district that became a trading center on the east-south border of west java province, especially the sale of agricultural products, where the traditional markets normally accommodate and sell agricultural products from surrounding as a new autonomous city, that is constructing massive infrastructure and public facilities, it can also be perceived that for the five-year period from 2007 to 2011 the construction and building sector provided a stronger portion to economic growth which declined in the latter five-year period (i.e., 2012-2017). another policy that is the focus of the local government is to give special attention to the village by providing financial assistance for infrastructure development and community empowerment, namely the provision of capital for village-owned enterprises. infrastructure development carried out by the village government is encouraged to be labor intensive, which means that every development implementation as far as possible involves the local community resources. the policy carried out by the banjar municipality government was implemented long before the central government also finally launched a †represents t values; **5 % confidence intervals; source: authors own elaboration 39 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 village assistance program. this policy is one of the positive effects of regional autonomy, because regions no longer build top-down but rather specific policies based on local wisdom and the most urgent needs according to each village in the region. however, the result of ssa indicated that the construction and building sector still maintain the aggregate grdp even though it was not considered a supporting sector of economic development. despite the small share in the gdp, the construction and building sector shared a high number of employed labor, which strengthens previous research from khan’s pakistan and sitsabo’s united kingdom studies (khan, 2008; dlamini, 2012). meanwhile, some nonreciprocal findings from south africa (dlamini, 2012) and turkey (berk & bicen, 2017) exhibit construction and building as an essential sector to economic growth. according to the ols regression analysis, local regulation and local road facilitation were two variables that contributed significantly and have a strong correlation to the economic growth during the ten-year period (i.e., p-values of 0.034 and 0.040, respectively). it can be inferred that local regulation offers positive value and effects to economic growth; in other words, the local government produces local regulatory precepts that give certainty to law and order (and in extension business confidence). one interesting policy adopted by the local authority is the postponement of the permit of establishment a new modern mini market since 2005, which has contributed to the survival of small retail stores. although this policy is relatively challenged by investors in the modern mini market, and quite influential on the small number of investors interested in investing in or opening a business in banjar, it has succeeded in encouraging the growth of small stalls and new small businesses. this situation considered as one of a good example of what this new autonomous city has done about how they maintain their economic activities by defending small entrepreneurs from the onslaught of big entrepreneurs. as such, regulation-based local knowledge also contributes significantly to the development of the city. in terms of local road facilitation, previous research has also detailed local road development as a good precedent to increasing economic growth via the movement of workers and market connections (asher & novosad, 2019). positive impacts from regulation are in line with research from the united states in which public policy is different in different areas, so that the service sector basically has an important role in the economy of this region jurisdictions. this affect has wide implications on business allocation and deregulatory effects on cutting the cost of production (fisher, 1997) and moving product. however, to enlarge a comprehensive understanding of the economic development in banjar municipality, we would stress the importance of conducting a socioeconomic survey that could elucidate any regulatory differences in terms of local, provincial, and national level benefits—and extend the research to reciprocal countries concerned with similar urban-economic phenomena. agus supriyadi, et al. sustainability policy in indonesia: case study … | 40 7. conslusion in the development economics it is generally perceived that efficiency of allocation means more productive resources while sustainable growth is more in line with economic, social, political, and institutional mechanisms via public and private services (todaro & smith,2012; paczoski et al., 2019). the economic structure, in terms of structural change theory, focuses on the transformation from traditional subsistence agriculture to a modern urban way of life (todaro & smith, 2012; russo & cirella, 2019). this study used grdp data with an underlying objective of determining what leading sectors contributed the most to economic growth since banjar municipality’s issuance as a new autonomous city. the following conclusions were obtained from 2007 to 2016. banjar municipality relies heavily on the following sectors: agriculture, livestock, forestry, and fishery; trade, accommodation, and food business; finance and business services; and other (i.e., public and private) services—in that order. thus, to accelerate economic growth, it is highly recommended these sectors continue to be properly facilitated within regional development and related policy plans. however, as noted with the exception of the mining and quarrying sector, the other sectors did not have a positive impact in terms of their competitiveness with other autonomous regions. in this regards, innovative local government development in creating policy that would increase the specialty in local products in the leading sectors should be emphasized. in terms of regional autonomy, the policy implementation process set in place by banjar municipality’s local government will need to continue to assess and develop policy in several public areas and facilities. local regulation, implemented in the region, should continue to focalize on protecting traditional market policy as well as lengthen, upgrade, and maintain road infrastructure. such regulation could include a community of practice approach (cirella et al., 2018; docherty, 2017) which would take into account local knowledge and consideration when piecing together city goals. if local policy expands local road infrastructure there would be an expectation of increased accessibility for the local inhabitants by way of mobility, doing business, and helping to protect the traditional market. new policies will need to be properly enforced and maintained for maximum economic success. in terms of sustainability, regional authorities will need to adjust and promote transitional steps to battering capacity, education, community knowhow, and policymaking skills (darmawan, 2008). hence, it is our recommendation, at this stage, that necessary central government intervention in terms of sustainability still be interplayed within the municipality’s policymaking outlook. 41 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 references akib, h., wagianto, e., daraba, d., & niswaty, r. (2019). shift-share analysis of the development of local competence as a basis for interregional cooperation in west sulawesi province, indonesia. journal of legal, ethical and regulatory issues, 22(1), 1-11. ali, m., egbetokun, a., & memon, m. h. (2018). human capital, social capabilities and economic growth. economies, 6(1), 2. andayani, k. (2018). economic base determination and influence of several variables against contributions percentage of the gdrp in aceh besar district. ms&e, 352(1), 012053. arent, d. j., tol, r. s., faust, e., hella, j. p., kumar, s., strzepek, k. m., ... & xu, h. (2015). key economic sectors and services. in climate change 2014 impacts, adaptation and vulnerability: part a: global and sectoral aspects , 659-708. asher, s., & novosad, p. (2019). rural roads and local economic development online appendix. asian development bank., “asian development bank support f or decentralization in indonesia.” special evaluation study, 2010. berk, n., & biçen, s. (2017, april). causality between the construction sector and gdp growth in emerging countries: the case of turkey. in 10th annual international conference on mediterranean studies, 10-13. bose, n., haque, m. e., & osborn, d. r. (2007). public expenditure and economic growth: a disaggregated analysis for developing countries. the manchester school, 75(5), 533-556. bps-statistic of banjar municipality. 2017, banjar municipality in figures. banjar mucicipality: bpsstatistic banjar municipality. cirella, g. t., iyalomhe, f. o., jensen, a., & akiyode, o. o. (2018). exploring community of practice in uganda’s public sector: environmental impact assessment case study. sustainability, 10(7), 2502. costanza, r., hart, m., talberth, j., & posner, s. (2009). beyond gdp: the need for new measures of progress. the pardee papers. crawley, a., beynon, m., & munday, m. (2013). making location quotients more relevant as a policy aid in regional spatial analysis. urban studies, 50(9), 18541869. daly, m. c., & regev, t. (2007). labor force participation and the prospects for us growth. frbsf economic letter. agus supriyadi, et al. sustainability policy in indonesia: case study … | 42 darmawan, r. e. d. (2008). the practices of decentralization in indonesia and its implication on local competitiveness (master's thesis, university of twente). dlamini, s. (2012, june). relationship of construction sector to economic growth. in international congress on construction management, canada. docherty, m. a. (2017). action research: a process for developing professional capacity within a community of practice. international journal of knowledge management studies, 8(1-2), 74-82. dogru, t., & sirakaya-turk, e. (2017). engines of tourism's growth: an examination of efficacy of shift-share regression analysis in south carolina. tourism management, 58, 205-214. drucker, j., kim, g., & weber, r. (2019). did incentives help municipalities recover from the great recession? evidence from midwestern cities. growth and change, 50(3), 894-925. dunn jr, e. s. (1960). a statistical and analytical technique for regional analysis. papers in regional science, 6(1), 97-112. endaryanto, t., firdaus, m., siregar, h., & budiman hakim, d. (2015). the impact of regional expansion on economic structure: a case study in lampung province, indonesia. international journal of sciences: basic and applied research (ijsbar), 23(2), 1-18. evans, g. k. (2008). spatial shift-share analysis of the leisure and hospitality sector on the gulf coast following hurricane katrina (no. 1368-2016-108539). fagerberg, j. (2000). technological progress, structural change and productivity growth: a comparative study. structural change and economic dynamics, 11(4), 393-411. fan, s., & chan-kang, c. (2005). road development, economic growth, and poverty reduction in china, 12. intl food policy res inst. feriyanto, n. (2014). determinant of gross regional domestic product (grdp) in yogyakarta special province. economic journal of emerging markets, 6(2), 131140. fernández, m. m., & menéndez, a. j. l. (2005). the spatial shift-share analysis-new developments and some findings for the spanish case. fisher, r. c. (1997). the effects of state and local public services on economic development. new england economic review, 53-66. głodowska, a., & pera, b. (2019). on the relationship between economic integration, business environment and real convergence: the experience of the cee countries. economies, 7(2), 54. 43 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 goschin, z. (2014). regional growth in romania after its accession to eu: a shift-share analysis approach. procedia economics and finance, 15, 169-175. greene, william h. econometric analysis, 7th edition. 7th ed. harlow: pearson education limited. 2012., 2. hu, w., & wang, r. (2019). which chinese cities are more inclusive and why?. cities, 86, 51-61. jahn, m. (2017). extending the flq formula: a location quotient-based interregional input–output framework. regional studies, 51(10), 1518-1529. jackson, r. w., & haynes, k. e. (2009). shift-share analysis. international encyclopedia of human geography, 125-131. johnston, a., & huggins, r. (2018). regional growth dynamics in the service sector: the determinants of employment change in uk regions, 1971–2005. growth and change, 49(1), 71-96. kelly, roy, “strengthening the revenue side.” in fiscal decentralization in indonesia a decade after big bang, edited by directorate general of fiscal balance ministry of finance republic of indonesia, 2012, 173–204. khan, r. a. (2008, august). role of construction sector in economic growth: empirical evidence from pakistan economy. in proceedings of the first international conference on construction in developing countries (iccidc), karachi, pakistan, 279-290. khusaini, m. (2015). a shift-share analysis on regional competitiveness-a case of banyuwangi district, east java, indonesia. procedia-social and behavioral sciences, 211, 738-744. kifli, f. w., & umami, a. (2017). contribution of economic sector and agricultural development to support food security in north sumatra province. advanced science letters, 23(7), 6428-6431. knudsen, d. c. (2000). shift-share analysis: further examination of models for the description of economic change. socio-economic planning sciences, 34(3), 177198. kowalski, e. (2000). determinants of economic growth in east asia: a linear regression model. honors projects, 74. kristiawan, k., ar, n. h., soemarno, s., & tyasmoro, s. y. (2016). leading product development of fruits using location quotient in tuban regency. iosr journal of agriculture and veterinary science, 9(8), 81–84. agus supriyadi, et al. sustainability policy in indonesia: case study … | 44 lankauskienė, t., & tvaronavičienė, m. (2013). economic sector performance and growth: contemporary approaches in the context of sustainable development. intellectual economics, 7(3), 355-374. liao, l., du, m., wang, b., & yu, y. (2019). the impact of educational investment on sustainable economic growth in guangdong, china: a cointegration and causality analysis. sustainability, 11(3), 766. liu, j., tang, j., zhou, b., & liang, z. (2018). the effect of governance quality on economic growth: based on china’s provincial panel data. economies, 6(4), 56. mallick, j. (2017). structural change and productivity growth in india and the people's republic of china. masaki, t. (2018). the impact of intergovernmental transfers on local revenue generation in sub-saharan africa: evidence from tanzania. world development, 106, 173-186. miller, m. a. (2013). decentralizing i ndonesian city spaces as new ‘centers’. international journal of urban and regional research, 37(3), 834-848. mogues, t., & benin, s. (2012). do external grants to district governments discourage own revenue generation? a look at local public finance dynamics in ghana. world development, 40(5), 1054-1067. moore, t., meck, s., & ebenhoh, j. (2006). an economic development toolbox: strategies and methods, 541. morrissey, k. (2014). producing regional production multipliers for irish marine sector policy: a location quotient approach. ocean & coastal management, 91, 5864. noisirifan, de fretes pieter, “analysis of leading sector (lq), structure of economy (shift share), and projection of grdp papua province 2018.” develop, 1(2) 2018. nosier, s., & el-karamani, a. (2018). the indirect effect of democracy on economic growth in the mena region (1990–2015). economies, 6(4), 61. oates, w. e. (1993). fiscal decentralization and economic development. national tax journal, 46(2), 237-243. paczoski, a., abebe, s. t., & cirella, g. t. (2019). debt and deficit growth rate reporting for post-communist european union member states. social sciences, 8(6), 173. paganelli, m. p. (2006). adam smith: why decentralized systems?. the adam smith review, 2, 201. 45 | international journal of informatics information system and computer engineering 2(1) (2021) 25 46 peng, c., song, m., & han, f. (2017). urban economic structure, technological externalities, and intensive land use in china. journal of cleaner production, 152, 47-62. riddington, g., gibson, h., & anderson, j. (2006). comparison of gravity model, survey and location quotient-based local area tables and multipliers. regional studies, 40(9), 1069-1081. rigby, d. l., & anderson, w. p. (1993). employment change, growth and productivity in canadian manufacturing: an extension and application of shift-share analysis. canadian journal of regional science, 16(1), 69-88. romyen, a., liu, j., sriboonchitta, s., cherdchom, p., & prommee, p. (2019). assessing regional economic performance in the southern thailand special economic zone using a vine-copar model. economies, 7(2), 30. russo, a., & cirella, g. t. (2019). edible urbanism 5.0. palgrave communications, 5(1), 19. sobari, a. (2011). “the impact of fiscal decentralization to the regional socio economic development in indonesia.” bogor agricultural university. bogor agricultural university. sondakh, l., & jones, g. (2003). an economic survey of northern sulawesi: turning weaknesses into strengths under regional autonomy. bulletin of indonesian economic studies, 39(3), 273-302. sukamdi. (2014). struggling out of recession: the influence of crisis on economic performance and welfare in java. bulletin of indonesian economic studies, 50(2), 295-296. supriyadi, a., wang, t., chu, s., ma, t., shaumirahman, r. g., & cirella, g. t. (2020). urbanization and population change: banjar municipality. in sustainable human–nature relations, 205-223. suryadarma, d., suryahadi, a., & sumarto, s. (2013). sectoral growth and job creation: evidence from indonesia. journal of international development, 25(4), 549-561. suryahadi, a., hadiwidjaja, g., & sumarto, s. (2012). economic growth and poverty reduction in indonesia before and after the asian financial crisis. bulletin of indonesian economic studies, 48(2), 209-227. suryani, g. (2015). regional financial autonomy and the economic growth in indonesia. bulletin of monetary economics and banking, 18(2), 1-22. todaro, m. p., & smith, s. c. (2012). economic development. edited by donna battista and david alexander. 12th ed. new jersey: pearson, 2015. wagner, a. (1890). finanzwissenchaft, winter, cf, leipzig. agus supriyadi, et al. sustainability policy in indonesia: case study … | 46 wilson, p., ping, t. s., & robinson, e. (2005). assessing singapore's export competitiveness through dynamic shift-share analysis. asean economic bulletin, 160-185. zhu, s., wu, q. j., & wang, y. (2011, september). impact of labor force on china's economic growth based on grey metabolic gm (1, 1) model. in proceedings of 2011 ieee international conference on grey systems and intelligent services , 262265. 1 | international journal of informatics information system and computer engineering 1 (2020) 47-78 a paradigm shift from optimal play to mental comfort: a perspective from the game refinement theory hiroyuki iida, mohd nor akmal khalid japan advanced institute of science and technology, 1 chome-1 asahidai, nomi, ishikawa 923-1211, jepang correspondence: e-mail: iida@jaist.ac.jp a b s t r a c t s a r t i c l e i n f o the game refinement theory focuses on the game designer perspective, where its application in various types of games provides evidence of the occurring paradigm shift. utilizing the logistical model of game outcome uncertainty, it provides a platform for incorporating gamified experience observed in games to be adopted in domains outside of game while retaining the context of the game. making games as a testbed, the implications of the game refinement theory have been observed in the educational and business perspective, while further explored its utility in interpreting some states of the human mind. in addition, a holistic view of design in games and in the real-world environments was discussed, where the prospects of the game refinement theory were also highlighted. article history: received 11 nov 2020 revised 20 nov 2020 accepted 25 nov 2020 available online 26 dec 2020 ble online 09 sep 2018 ___________________ keywords: index terms—game refinement theory, gamification, edutainment, momentum-in-mind, comfort-in-mind, serious game. international journal of informatics, information system and computer engineering international journal of informatics information system and computer engineering 1 (2020) 47-78 h iida and m n a khalid a paradigm shift from optimal...| 48 1. introduction originated from the idea of the existence of mixed strategy equilibria in two-person zero-sum games (neumann, j. v., 1928), game theory (gt) had been widely utilized as a useful tool in many fields such as economics, political science, psychology, logic and biology. however, the gt focuses on the perspective of the player’s side. looking from the perspective of the designer’s side, the game refinement theory (gr) is a unique theory derived from the classical game theory. the gr is defined as the measure of attractiveness and the sophistication of games based on the game outcome uncertainty (panumate et al., 2015). previous works such as (diah, et al., 2014), (xiong, s et al., 2014), (xiong, s., and iida, h, 2014), (panumate et al, 2015), focused on various spectrum of games such as sports, board games and video games. it shows that the game refinement values of those popular games support the previous assumptions of a balanced range of game refinement value which is gr ∈ [0.07,0.08] (xiong, s et al., 2014). while many research works have been done using gr that focuses on game entertainment of many gamerelated domains, there had been limited applications to the real-world settings. the goal of this paper is to highlight the current change of paradigm in game design from the lens of the gr framework, where the shift of focus from making an optimal play to providing mental comfort has been observed from recent works. the organization of the paper is as follows. the game refinement theory is described in details from the perspective of continuous game (such as sport games), discrete game (turn-based and board games), and mixture of both perspectives (section ii). then, its application in the context of education and business are investigated where gr theory is applied in relation to current concept of edutainment and gamification in the real-world settings (section iii). next, in-depth interpretation of the gr theory with respect to the perceived nature of human mind, specifically on understanding the sense of curiosity and boredom, is investigated (section iv). in addition, general discussion on the overall perspective of the gr theory is given in section v while section vi concludes the paper. 2. game refinement theory a personal decision process in mind typically involves considering all possibilities of available options which then are reduced to only a single decisive solution. this process commonly involves both the aspects of skill and chance. in a case of a board game, an experienced player may identify only a few decent moves out of all possible instances (iida h., 2008). however, only one solution has to be decided as the final solution. this idea has been expanded to establish the mass-in-mind model, which being integrated into the game refinement model. fig. 1 presents the generic selection process in a player’s mind. fig. 1. the player decision selection process while all possible instances are relatively large, some of them are out of an experienced player’s consideration 49 | international journal of informatics information system and computer engineering 1 (2020) 47-78 as they might lead to deficiency or failure. the effective branching factor b of a player is a subset of the branching factor, in which only acceptable solutions are concerned. effective branching factor of a player is the number of instances, which are satisfyingly perceived by that player in a single decision. reduction rate of branching factor (say α) for the minimax-based game tree where b and b (1 ≤ b ≤ b) is the average branching factor and the effective branching factor, respectively. thus, α is defined by (1). the effective branching factor b is significantly smaller than the branching factor b but not underneath 1, so 1 ≤ b ≤ b. (1) a. a game progress model the general game refinement model was proposed based on the concept of game information progress, where the gap between boardgames and sports games was bridged (sutiono et al., 2014, iida, h., 2018). the “game progress” can be one of two things. one of it is the game speed or scoring rate. meanwhile, another can be the game information progress which focuses on the game outcome. game information progress presents the degree of certainty of the game’s results in time or in steps. having the complete information on the game progression (i.e. after its conclusion), the game progress x(t) will be given as a linear function of time t with 0 ≤ t ≤ tk and 0 ≤ x(t) ≤ x(tk), as shown in (2). (2) here, tk and x(tk) corresponds to the end of task and the goal to achieve in task, respectively. (2) implies that goal of the task under consideration must be achieved within the time tk, meanwhile it may not be shorter time than tk (iida h., 2018). under the given time constraint (fig. 2) that task should be completed within the time t with 0 < t < tk where x(t) = x(tk). however, the game information progress given by (2) is unknown during the in-game period since t < tk. the presence of uncertainty during the task, often until the final moment of the task, would cause the task progress to be exponential. hence, a more realistic model of information progress of a given task is given by (3). fig. 2. an illustration of normal activity becomes gamified under a time constraint (adapted from (iida h., 2018).) (3) here n ∈ ir stands for a parameter related to the degree of difficulty of completing the task. in the context of (3), lower n implies a very boring task while higher n would imply higher difficulty of completing the task. by deriving the task’s information h iida and m n a khalid a paradigm shift from optimal...| 50 progress, when n ≥ 1, the “velocity” in the sense of information dynamics can be obtained in (4). (4) when n ≥ 2, the second derivative of the task’s information progress or “acceleration” in the sense of information dynamics, which obtainable by deriving (3) twice. then, solving it at t = tk, (5) was obtained. the game information progress modelled in (5) was assumed to be the abstraction of the information in a human brain. as the physics of information in the brain is yet to be discovered, it is likely that the acceleration of information progress is subjected to the forces and laws of physics. too little game information acceleration may be trivial for human observers and players to compute. in contrast, too much game information acceleration would cause frustration, overwhelming and even incomprehensible to humans. representing x(tk) and tk as g and t respectively, the gr formulation in its root square is given in (6). therefore, increases in the value gr is expected to increase the excitement of the game, due in part to the uncertainty of game outcome. application of the gr measure in popular sports is given in table 1. (6) table 1. the gr measure for popular sport games (takeuchi j., 2014, ramadhan et al.,2015, panumate, et al., 2017). games 𝑥(𝑡𝑘) 𝑡𝑘 𝐺𝑅 basketball 36.38 82.01 0.073 soccer 2.64 22.00 0.073 volleyball (side-out: 15pts) 15 52.52 0.121 volleyball (rally point: 30pts) 30 53 0.104 volleyball (rally point: 25pts) 25 44 0.114 table tennis (pre-2000) 57.87 101.53 0.075 table tennis (post-2000) 54.86 96.47 0.077 badminton (side-out) 30.07 45.15 0.121 badminton (current) 46.34 79.34 0.086 baseball (major league) 4.57 34.23 0.062 b. a board game model for a board game, the definition of branching factor is the amount of all possible instances in a single decision while the game length is the number of steps from the beginning to the ending or the resignation. a game tree is constructed by recursively attaching all possible transitions to the initial position (xiong, s et al., 2014) let b and d be the average branching factor and the average game length respectively. a single decision can be illustrated in fig. 3. by considering the geometry, the distance d is obtained by according to the pythagorean theorem (maor e., 2019). however, one is much smaller than b and left from the consideration. 51 | international journal of informatics information system and computer engineering 1 (2020) 47-78 hence, the distance d becomes . assuming the outcome continuously becomes evident, the model of game progress x(t) is determined by the proportion of the d and the game length d. therefore, the model of game progress becomes . in general, (7) was obtained. fig. 3. a generic single decision in a game tree the value of c and n(n − 1) in (5) had been discussed in previous works (xiong et al., 2017) following the game progress model, the uncertainty of the outcome renders the x(t) exponential. therefore, the more realistic model of game information progress becomes (8). then, similarly in obtaining the game refinement measure gr of the game progress model, the gr of the board game model is obtained by deriving (8) twice and taking its root square form, as shown in (9). application of the gr measure in popular board games and other multiplayer games is given in table 2. if the same model were also extended to the domain of the video games, then the b and d corresponds to the average number of successful attempts and average number of attempts per game, respectively. as such, the gr measure of some popular video games is also provided (table 3). (8) (9) table 2. the gr measure for popular board and other multiplayer games (iida et al., 2004, xiong et al., 2014, panumate et al., 2016, panumate et al., 2017) games x(tk) or b tk or d gr uno 0.976 12.684 0.078 roshambo 3 1.5 1.155 ufo catcher 0.967 13.30 0.074 western chess 35 80 0.074 chinese chess (xiangqi) 38 95 0.065 japanese chess (shogi) 80 115 0.078 go 250 208 0.076 mah jong 10.10 39.56 0.078 table 3. the gr measure for popular video games (xiong et al., 2014, panumate et al., 2016, (punyawee et al, 2016, xiong et al., 2017) games x(tk) or b tk or d gr dota v6.48 69.2 110.8 0.075 dota v6.64 68.4 110.4 0.075 dota v6.80 68.6 106.2 0.078 snake ai (wall) 4.44 34.98 0.060 snake ai (no wall) 2.18 17.03 0.087 starcraft 2 (terran) 1.64 16 0.081 starcraft 2 (protoss) 1.55 18 0.069 starcraft 2 (zerg) 1.61 20 0.082 king of fighters 98 14.6 36.7 0.104 h iida and m n a khalid a paradigm shift from optimal...| 52 c. a swing model the game progress model supports only a game with uniformed scoring rate, e.g., soccer. in soccer, a successful shoot is particularly challenging to obtain and regarded as one score. supposing that the gained score is multiplied, the measure of game progress model may alter, but the real essence of the game remains unchanged (kananat et al., 2016). however, scrabble players earn several marks in turn. this incident happens in a case of the games with the non-uniformed scoring system. therefore, scrabble has nonuniformed scoring rate and not directly compatible with the game progress model. instead, the swing model is introduced by defining swing turnover as a state transition in mind during the game progress among some possible states. in a game with non-uniformed scoring rate, the average amount of swing turnover s is proposed as a measure for counting the actual successful shoot. although the transition among possible states may differ for a different domain, we consider two cases: advantage and disadvantage. it is rumored that maximizing own profit while minimizing others have been the general principle of the intelligent decision maker (harsanyi et al., 1988). obtaining the highest score is the goal of playing scrabble. in each step, players are taking their turn to attempt to have the advantage over the opponent, which is considered as the actual successful shoot g, if successful. let d0 be the average turn that player potentially turns the swing. due to difficulty in measuring d0, the game length d is used as its approximation. the game refinement gr of the swing model is obtained by (10). (10) swing model is an approximation of game progress model as an exciting game when there is a proper amount of swing turnover as opposed to a singlesided game (kananat et al., 2016). the game refinement measure reflects the balance between player’s strength and surrounding randomness in a game considered (xiong et al., 2017) while a superior value implies that a chance becomes a stronger factor, a game with an extreme game refinement measure might overwhelm a player with the information, which results in frustration. as such, the gr measure based on (10) for scrabble based on 20,000 match records from crosstable.com is provided (table 4). table 4. game refinement value of scrabble game average number of swing turnovers (s) 5.096 average number of total rounds (d) 27.145 total match record data 23,164 game refinement value (gr) 0.083 3. game refinement theory meets the real-world in a real-world setting, design of game elements is highly dependent on the context of interest. incorporating 53 | international journal of informatics information system and computer engineering 1 (2020) 47-78 game elements in education leads to an innovative teaching-learning methodology known as “edutainment”, which involves the mixture of entertainment and education or marriage of education with entertainment (aksakal n., 2015, agarwal s et al., 2019). such methodology may be inclusive of different levels of interactions, motivations, experiences, and mental acceptance in the educational process through elements perceived to be entertaining, commonly found in games. the players of such games benefited from collaborating among themselves both inside the amusement condition (e.g. through networking on the web-based groups) and around it (e.g. through sharing game-related data and assets) (agarwal s et al., 2019). in a broader sense, adopting game design elements in nongame contexts is a popular concept known as “gamification”. the concept of “gamification” was first launched in 2008 in a blog post by brett terrill (terrill, b. 2008). he described gamification as taking the game mechanics and applying them to other web properties to increase engagement (huotari, k., and hamari, j., 2012), which becomes a widespread industry use (deterding, s et al., 2011). gamification can be defined as adopting game design elements in non-game contexts, without changing the existing practice of that context, in order to improve user experiences in term of engagement and learning, employee recruitment and evaluation, ease of use and encourage physical exercise (huotari, k., and hamari, j., 2012, huynh, d et al., 2016). gamification also enhances the player’s ability to understand a specific domain of studies such as language learning or music. although there can be many methods to edutainment and gamification, the gr theory provides an alternative method to both concepts from the educational and business standpoint. note that the gr theory is adopted with only minor changes on defining the context of the “game progress” with respect to the application domain. the gr value still retains its original intention, which is the measures of attractive and sophistication in game design. a. educational perspective application of the gr measure with respect to educational perspective had seen some potentials, which demonstrated in two type of edutainment games: mindsnacks and duolingo3. these games were selected due to several reasons. firstly, both of the games were language learning platforms. secondly, they shared similarity in term of the milestone technique used as part of the game elements in the learning process. thirdly, both of them adopted learning methods that measurable in number (lessons or tasks). mindsnacks are specialized games for individual languages including chinese, japanese and several european languages. it spreads learning lessons such as vocabulary, spelling, and grammar across nine different mini-games with their own h iida and m n a khalid a paradigm shift from optimal...| 54 hand-drawn art. a user is first given three available minigames to start while the rest is unlocked when the user reaches a certain number of levels for each. by adopting the milestone technique used by game designer, there is a certain number of levels that require completion (by performing several tasks) before unlocking the mini-game in a particular milestone (izzyann, f et al., 2018). therefore, lv and t is the average number of levels required to unlock the mini-games and the average number of tasks, respectively. to calculate the gr value, with reference to (6), let g and t equals to lv and t, respectively, which corresponds to the progress model for mindsnacks. meanwhile, duolingo is a publicly available gamified free languagelearning platform since 2012 which is created by professor luis von ahn and his graduate student severin hacker. as of april 2016, it offers 59 different language courses across 23 languages. duolingo is a product of gamification, where game elements are utilized to create enjoyment in the learning process. in a language course, the goal is to complete a study by getting all badges in the skill-tree. to achieve a badge, users must complete all lessons in a skill. as such, the game progress of duolingo language course can be measured by skills. let s and l be the average number of skills and the average number of lessons in the same language courses, respectively. therefore, with reference to (6), the s and l in duolingo are equal to the g and t, respectively, in order to calculate its gr value. a comparison between mindsnacks and duolingo on japanese language course using the gr measures is given in table v. referring to the previous work, the spanish language course of duolingo falls onto the range of gr ∈ [0.020,0.026] which is reasonable for the serious-like game environment (huynh, d et al., 2016). similar tendency occurred in both platforms for the japanese language course where the gr value for each milestone decreases gradually. also, there is a big gap between grvalue at the first milestone of both platforms due to the big difference of the average number of tasks or lessons needed to complete it. this difference in value indicates that duolingo is much more attractive to beginner learner while mindsnacks requires learner with strong motivation since it’s much difficult to beginner. table 5. game refinement measures of mindsnacks and duolingo on the japanese language course milestone lv t gr milestone s l gr totems 3 5 0.35 1 7 34 0.08 blooms 6 17 0.14 2 15 66 0.06 stacks 10 37 0.09 3 27 130 0.04 bubbler 15 85 0.05 4 40 185 0.03 dam builder 21 141 0.03 lv: average number of levels required to unlock the mini-games; s: average number of skills; t/l: average number of tasks/lessons; the gr value for both mindsnacks and duolingo started to decline in the 55 | international journal of informatics information system and computer engineering 1 (2020) 47-78 subsequent milestones as the learner progresses, where both resulted in the same gr value of 0.03 at the fifth and fourth milestone, respectively. additionally, mindsnacks has one extra milestone compared to duolingo, making mindsnacks much more appealing towards serious but casual learners. in contrast, duolingo may appeal to additional game elements (such as winning streaks (huynh, d., and iida, h., 2017) to retain learner’s motivation since only small sense of achievement can be elicited (from the small changes of the gr value between milestones). as such, a longer milestone may effect users’ feeling of achievement. finally, observing the accumulated tasks or lessons at the end of each milestone, mindsnacks had a lower t value compared to l value of duolingo. this means the enjoyment and motivation of finishing the tasks or lessons inversely proportional to its number. b. business perspective gamification employs game design elements which are used in non-game contexts to actively improve the user’s engagement by applying game-based thinking through badges, rewards programs and points (zuo, l., and iida, h., 2017). there are three main ingredients (but not limited to) that offer gamification in various real-world application; game element, game design, and non-game context. game elements involves building some form of service which utilized usage of bits and pieces of games (i.e. interactive graphical user interface) to improve the user experience. meanwhile, game design is the inclusion of game playing feeling through game elements such as the points and tiers system which organized systematically and encourage enjoyment in such activity. finally, non-game context can be translated to anything other than the game in which the objective is outside of the game; thus, related to their needs or business. with respect to the gr theory application and gamification in the business perspective, two types of business strategies were analyzed from the customer’s point-of-view; sales promotion and loyalty programs. 1) “discount” sales promotion: the money saved by the sales promotion ‘discount’ is considered beneficial to the customers. this is because customers have little knowledge and interest about the cost of a product of interest (zuo, l., and iida, h., 2017). however, they would care more about the price of the product (i.e., how much money can be saved by the discounting). if the rate of discount is reasonable from the perspective of customers, the “game of discount” would have a comfortable zone value of game refinement, similar like the sophistication of the boardgames and sports. as such, the money saved by the sales promotion ‘discount’ is considered as the benefit for customers where the total benefit of the customer’s shopping activity in a certain time is measurable. based on such information, the game progress model can be constructed by two factors: the total benefit of a certain product (g) and the total normal price without discount (t). important note is that discount in the sales promotion h iida and m n a khalid a paradigm shift from optimal...| 56 typically expressed as percentage. to normalize the model in (6), we assume that the benefit of a product is ranged from 0 to 100 and the normal price is 100. then, the model in (6) is extended to (11) that gets the grc of different extent of discount for customers where d stands for the discount coefficient (i.e. 10% discount equals to 0.9 discount coefficient) (zuo, l., and iida, h., 2017). (11) it is important to identify a reasonable discount zone for customers. it is expected that lower gr would be less attractive for customers, whereas higher gr would be of some concern about quality and hard for the sustainability of seller. two data sets were collected for the analysis. the first data set is intended to optimize the seller’s total profit within a time frame using computer simulation in order to maximize total sales volume from the seller’s standpoint. the second data set is collected from bestsellers (bento boxes & water bottles4) of a popular on-line retailers (amazon jp5) which corresponds to the customer’s standpoint. then, both data set was set with a fictitious market demand of 10,000 products within a year. the total amount sale is determined by two factors: quantity and price. the discount coefficient will be changed every week and range from 0 to 1. the demand of product follows the fdistribution and the customers follow the poisson distribution. tables vi showed the discount rate by simulation and real data of amazon jp where the discount rate of simulation is ranged from 21% off to 39% off, whereas the top sales product of amazon jp is ranged from 42% off to 59% off. note that the grc results of simulation is best for the seller but considers little about customers. however, grc of amazon jp is ranged from 0.065 to 0.077, which is located in the zone value (gr ∈ [0.07,0.08]). thus, under the assumption that the sales promotion ‘discount’ is a fascinating game, the simulation concerns about sustainability and amazon jp concerns about popularity of the game. table 6. discount rate and measures of game refinement (grc) for simulation, amazon jp, and recommend discount rate (%) grc simulation 42 59 0.065 – 0.077 amazon jp 21 39 0.046 – 0.062 recommended 49 64 0.07 – 0.08 the attractiveness of games often comes within the approximate zone values (gr ∈ [0.07,0.08]) where its corresponding discount zone is ranged from 49% to 64% off (tables 6). the essence of game refinement theory is to find a comfortable acceleration (the sense of thrill) for game players (or customers). hence, the discount rate should not be too high. for example, too high grc may also cause some concern such as quality of a product of interest. this is why some shopping festivals, like “black friday” in america and “1111 shopping festival” in china, set their discount rate 50% off since this discount rate falls on the comfortable 57 | international journal of informatics information system and computer engineering 1 (2020) 47-78 zone. higher grc means higher entertainment impact and more popular which would be able to motivate customers to buy more. the essential idea is to offer an appropriate discount zone based on the previous knowledge of the gr theory where benefits of game thinking (such as discounting) can be adopted in the business domain. 2) customer loyalty program: in today highly competitive marketplace, customer retention is an important factor for the success of businesses because the cost of finding a new customer is about 5 to 10 times more than to keep the existing one. customer retention is defined as the activities and actions of companies and organizations to reduce the number of customer defections (xin, o et al., 2017). loyal customers tend to spend more (67% more than new customers) and often know exactly what they want when purchasing from a certain brand. unfortunately, companies tend to lose around 13% of their customers every 5 years (kumar, k., 2016). therefore, loyalty of a customer is critical to the sustainability of a business. generally, loyalty program refers to a reward program offered by a business to customers who frequently make purchases (xin, o et al., 2017). loyalty program typically requires customers to register with their information and customers will be given a unique membership id or physical membership card to be used when making purchases. the business owner can track behavior and history of their customer purchase in order to recognize the loyal customers and hence reward them. meanwhile, customer engagement is further improved when the feeling of appreciation by the company is established with the customer. as such, three case studies of the customer loyalty program are analyzed using the gr theory; starbucks loyalty program, hotel loyalty program, and frequent flyer program (ffp). starbucks is a fast growing coffee shop business with 7% annual growth rate (xin, o et al., 2018). according to statistic in 2016, there is increase of 26% in profit and 11% total revenue when starbucks introduced the “my starbucks rewards” program (fisher, e., 2016). “my starbucks rewards” is a free loyalty program that gives exclusive member offers and allow customers to earn rewards (such as free drinks, foods and refills). customers simply need to pay for any starbucks product with a registered starbucks card to earn a reward. each time a purchase is made, customer will earn a specific amount of stars that can be redeemed for free starbucks treat. the terms and conditions of “my starbucks rewards” varies according to different countries. as of 2016, the number of starbucks licensed stores globally is ranked by unites states (us) as the first with 5,292 stores and followed by china with a total of 1,110 stores (xin, o et al., 2017). in order to calculate the gr value based on (6), the number of free items redeemed and total number of items consumed is represented by g and t, respectively. the evolution of “my starbucks rewards” since 2009, starbucks us had continuously made minor changes on the rules in redeeming free items (assuming the average price h iida and m n a khalid a paradigm shift from optimal...| 58 of a cup of starbucks coffee which is $4 and rmb27 in us and china, respectively). hence, the gr value of different versions of “my starbucks rewards” from the perspective of point system in gold level is analyzed (table vii). table 7. differences of “my starbucks rewards” in us and china and gr values of starbucks loyalty program (point system) in earning a free item in gold level based on versions country us china levels 2 3 req. to reach/stay (gold level) 38 47 req. to get 1 free item (gold level) 16 10 year version g t gr 2009 visit-based 1 15 0.067 2012 visit-based 1 12 0.083 2016 spend-based 1 16 0.063 req.: purchase requirement in number of cups per customer from 2009 to 2012, changes in “my starbucks rewards” rules increased the gr value. this means that customers in gold level are required to spend less for a free item redemption. on the other hand, comparing the visit-based in 2012 with the spending-based in 2016 of "my starbucks rewards", the gr values decreased from 0.083 to 0.063. this result may be explained by the fact that starbucks requires its customers to spend more in order to get a free item. before introducing the spending-based reward, customer can spend less than $4 (the average price of a cup of starbucks 1 all company names, loyalty rewards names, trademarks, and pictures are properties of their respective owners. coffee in 12 transactions) to earn a free item. since starbucks values its loyal and high-spending customer, the reward rules change so that star is earned based on amount of spending. these changes are less entertaining and more challenging for the customers whom spend minimum amount to earn stars or free items. in addition, it was found that the requirement for a customer to reach or stay in gold level in us is lower than in china. however, the requirement for a customer to get one free item in gold level is higher in us as compared to china. another customer loyalty program, called the hotel loyalty program, involves enticing frequent hotel guests or other business travelers to favor that particular brand or group of hotels over others when selecting a hotel (zuo, l et al., 2018). the world’s largest hotel chains, including marriott, starwood, hyatt, hilton, ihg (intercontinental) and others, all operate reward programs1. hotel loyalty program usually adopted tiers system (up to four tiers) and point system (point accumulation). a newly joined customer has several benefits, such as the complimentary inroom internet access and the guarantee of the lowest rate to encourage retention in the loyalty program. joining the scheme and staying nights in a hotel will typically accumulate points where it can be redeemed for free nights in hotels (zuo, l et al., 2018). higher levels in the scheme will generally offer a combination of benefits 59 | international journal of informatics information system and computer engineering 1 (2020) 47-78 (i.e. complimentary room upgrades, bonus points and hotel’s amenities such as conference room, lounge and other privileges). an elite status is typically reached by staying a certain number of nights say 60 or more during the membership year. the two mechanics (tiers and points) seem to be irrelevant, however, when looking back to the whole structure of the hotel loyalty program since it was found that the two mechanics are correlated with each other. the elite member could accumulate their points much faster with the additional points policy which also means that they could get more free nights (zuo, l et al., 2017). to calculate the gr value for hotel loyalty program of a tier system, the number of successful qualifying nights corresponds to the g and the total number of days within a year (365 days) corresponds to the t, where the result is given in table 8. basically, each hotel has almost the same strategy of membership management and tiers. however, marriott hotel has a higher requirement for the silver status. it means that marriott requires customers to spend more on its loyalty program. by observing the gr values, we may conclude that the gr value between 0.000 (become a member) to 0.027 (platinum level) would be the activity zone for the hotel loyalty program. this means that a customer can enjoy much more only when they becomes the elite status membership. thus, the hotel loyalty program may be recognized as a kind of skill-based games. table 8. gr measures for hotel loyalty program with focus on qualifying nights: five hotel groups compared hotel groups gr silver gold platinum ihg 0.009 0.017 0.024 hilton 0.009 0.017 0.021 starwood 0.009 0.019 0.027 marriott 0.014 0.019 0.024 hyatt 0.009 0.015 0.021 meanwhile, the points for redeeming free nights dramatically differ from the location and brands, which highlights the consideration of hotel location when applying the gr measure to different cities. thus, three big cities were chosen: shanghai, tokyo, and paris. in addition, as the gr measure requires the highest level (skillful player) to make the result more objective, the most senior membership was taken as the sample to calculate the gr value. as such, each member would be playing the same game. as such, the number of redeemable free nights (g) and the number of successful qualifying nights (t) is considered for the point system (table ix). the results showed that the average price and redeem points differ based on location. however, ihg and hilton have the same gr values regardless of the location which implies that these two hotel groups have the same strategy when dealing with the free nights. the gr value of starwood and hyatt slightly vary with the location. on the other hand, the result of marriott greatly differs in location. also, both starwood and marriott provide a higher game element for shanghai h iida and m n a khalid a paradigm shift from optimal...| 60 location; thus, showing that marriott is dedicating its resources in china location by providing a more enjoyable experience in shanghai. another customer loyalty program, called the frequent-flyer program or ffp, is a loyalty program offered by airlines. from a historical perspective, ffp is considered to be the world’s largest gamified service (zuo, l et al., 2017). many airlines have ffps designed to encourage customers to accumulate points (also called miles or segments) that redeemable for air travel or other benefits travelers can receive from the program (zuo, l et al., 2017). the ffp first started with the “advantage” program which was characterized by a series of inventions that improved airlines’ revenue streams, increased customer recognition, reward customers for using the airline, and to promote future customer loyalty (de boer, e. r., and gudmundsson, s. v., 2012). gamification of the ffp leverages the intrinsic human motivations of accumulating rewards or miles and build up the users’ motivation while ensuring the engagement is continuous (zuo, l et al., 2017). it can be defined as a service quality attribute that consists of two systems: tiers system and points system. the tiers system is commonly represented by four levels: member, silver, gold and platinum. the tiers system is an application of progression levels or difficulty levels just as in video games. depending on the airline’s regulations, different statuses enjoy different levels of rewards, additional points, priority check-in and lounge access. the membership requirement of tiers system of four chinese ffps is given in table x. meanwhile, the points system (or miles) is a type of virtual currency used predominantly within the game world (zuo, l et al., 2018). as a new player, one is offered a “qualifying segment” by the system which allows them to get fast bonuspoint earnings; thus, instant gratification. most, if not all, programs award bonus earnings to premiumcabin passengers and their elitestatus members based on tier status. a common bonus for these passengers is to earn an extra 15%–55% of the miles flown on a given flight. points can be redeemed for air travel, other goods or services, class upgrades and airport lounge access. for the tier system, the game progress model of a ffps can be determined based on the action of qualifying segments of membership tiers, which defined by the number of successful qualifying segments g and the total number of segments within a year t (365 days). here, we consider the normalized model that the customer can usually get one segment in one day (table 11). for the point system, data is collected by considering the flight distance. as game refinement requires the highest level (corresponding to the skillful player) to make the result more objective, only the most senior membership was used as a sample. since the higher status of the customers, the more miles/points they obtain. however, the points for redeeming free segments differ dramatically from the distance. there are three kinds of free segments considered: short, medium and long. thus, the impact of a free 61 | international journal of informatics information system and computer engineering 1 (2020) 47-78 segment for the highest membership of fortune wings club and the distance within a year to illustrate game sophistication and game experience. the total points one can earn within a year with considering the 80 segments (xiy-pvg) of the domestic travels with the point bonus are about 170,000. then, the data can be collected from the official website to check the points required for a free flight7, where the results is given in table 12. from table 11, every airline company has almost the same strategy of membership management composed of four tiers, except sky pearl club that excludes the platinum tier. maintaining or promoting the tier corresponds to overcoming a challenge in a game. the trend of these four ffps increases the gr value with the tier promotion, implying that the tiers system is offering a fun-game experience. this leads to a winwin scenario: customers improve their loyalty and companies increase their revenues (zuo, l et al., 2018). one dimension is the tiers system involving the sustainability of the frequent customer, while another is the profitability of points system which indicates the popularity of the gamified service since profit may encourage more customers involvement. meanwhile, table 12 shows that the highest gr value is 0.045. it is important to note that the more free segments is redeemed, the less challenge is met. this implies that the points system offers a serious-game experience. the points system concerns about the benefit for customers who may feel serious-game experience (a good balance between customer’s capacity and challenge) but related to the popularity. the free ticket is so compelling as everyone. table 9. gr measures for hotel loyalty program in shanghai/tokyo/paris with focus on free nights: five hotel groups compared hotel groups ave. price ($) ave. redeem points free nights gr ihg 75/208/148 18421//50000/34778 6/6/6 0.033/0.033/0.033 hilton 125/190/190 31400/55400/55000 3/3/3 0.029/0.029/0.029 starwood 115/247/335 6900/16000/19769 12/10/10 0.035/0.032/0.032 marriott 143/208/285 13026/36250/32800 15/7/10 0.052/0.035/0.042 hyatt 146/336/233 13285/20667/18000 4/6/51 0.033/0.040/0.037 4. game refinement and the perceived nature of mind game has existed since 2600 bc and has been the primary conditions of human culture and cultivated various complex activities of the human society such organization, languages, philosophies, war, and art (huizinga j., 2014). in the beginning, the game was just for relaxation after a busy work or making a simple decision. with the development of society, the game process becomes more complex and meaningful. h iida and m n a khalid a paradigm shift from optimal...| 62 table 10. tier membership levels based on four frequent-flyers programs membership tier miles/segments extra bonus (%) eastern miles silver 40,000/25 15 gold 80,000/40 30 platinum 160,000/90 50 phoenix mile silver 40,000/25 25 gold 80,000/40 30 platinum 160,000/90 50 fortune wings club silver 30,000/20 25 gold 50,000/40 50 platinum 100,000/80 55 sky pearl club silver 40,000/20 15 gold 80,000/40 30 table 11. gr measures for tiers system from four ffps membership tier gr eastren miles phoenix miles sky pearl club fortune wing club silver 0.014 0.014 0.012 0.012 gold 0.017 0.017 0.017 0.017 platinum 0.026 0.026 0.024 desires to enjoy a flight without payment as well as highlevel services. thus, these two gamified services organized systematically is to create a program to maintain customer loyalty. table 12. gr measures for points system in fortune wings club segment point free qualifying gr req. flights segment gr short (xiy-pvg) 13 13 80 0.045 medium (xiy-nrt) 28 6 80 0.031 long (xiy-cdg) 45 3 80 0.022 req.: point requirement ×103 63 | international journal of informatics information system and computer engineering 1 (2020) 47-78 historically, harnessing the power of play had been attempted to serve the purpose which is beyond entertainment. game designers are constantly working on stronger ai players to compete against human or other ai champions. similarly, there are human society constantly searching for ways to effectively disseminate ideas, educators constantly seeking appealing learning process, and companies constantly seeking to engage their customers (gino, f., 2018). the key factor to human conscience that triggers learning is curiosity, which observable from both entertainment and educational perspectives. although a mechanical definition of curiosity is unknown, the perceived curiosity from the learning process is measurable from the development of edutainment game. in most cases, edutainment game design tends to favor either extremes (entertainment or education), but not both. similarly, online learning games (such as the login system) that typically involved monotonous and repetitive action, tend to prioritize either the feeling of attractiveness (i.e. reward) or experience of engagement (i.e. loyalty program), but not both. due to aforementioned situations, the perceived curiosity are modelled in two distinct form that corresponds to the player’s state of mind: (a) by the growth rate of learning; and (b) the diminishing rate of engagement. the former is named as the momentum-in-mind while the latter is named as comfort-in-mind. note that, up to this point, the usage of gr value provides a very different interpretation to game design, which is highly dependent on the context of interest. a. momentum-in-mind corresponding to the perceived “curiosity” of the player when playing a game, the momentum-in-mind is defined as the growth rate of mind of the player while moving from one level to another by acquiring a set of information from each level (agarwal, s et al., 2019). in the context of edutainment, momentum is required to be maintained relatively with the time, because curiosity triggers increases the activity in the brain. this helps us identify the perfect combination of education and entertainment in games. as such, the proposed momentuminmind (p~) is formulated as (12). here, the t is the time taken to acquire information to reach a higher level in the game. it can be measured as the difference in the levels or in the unit of time. (12) table 13 summarizes the mathematical model of game progress using momentum. if the time is reduced, momentum increases and vice versa. thus, momentum is inversely proportional to time if the progress (h/l) is kept constant. in the context of the board game, the growth is not continuous and mostly redundant. as such, the momentum-in-mind is obtained by (13). b is the information of the game or game progress which the player learns actively while playing the game and t is the time for the player to acquire that information. h iida and m n a khalid a paradigm shift from optimal...| 64 table 13. momentum-in-mind notation game progress model h game information at higher level l game information at lower level t time taken between the levels p~ = (ln h) · 1 l t momentum (p~) acquired (13) 1) game play stages: it is supposed that most player in games proceed in the manner of two stages, called the ‘skill acquiring’ stage and ‘competitive’ stage. the skill acquiring stage is the stage when the player internalizes the skill required for the game (learning). as such, the competition is usually with oneself. at this stage, the goal is set against oneself and that goal becomes the ‘mental’ barrier (or ‘inertial’ level) that requires to gain enough momentum. as such, the momentum is expected to be low which implies high learning experience. another stage is called the competitive stage where the player starts to compete with other players (and get better at it). since competing with other player involved some level of skill and chance. the uncertainty would become higher which implies high sense of excitement and enjoyment. most edutainment games lack this stage where very minimal competition existed. therefore, the learning growth during this period is very low while the momentum is high. 2) application to competitive games: several examples of sport games include table tennis, basketball, soccer, badminton, and so on, while example of board games are go, shogi, chess, baghchal, and so on. based on the previous work by (panumate, c et al., 2017), the branching factor b and the game length d were taken where the gr value were calculated by using (6) and (9), while the momentum p~ were calculated by using (13), where d is considered as the length t for calculating the momentum for the aforementioned sports and board games (table 14). based on the calculated p~ and gr values, similar pattern was observed where increase in the gr also increases the p~, and vice versa. this aligns with current understanding of the game play stages where the considered games only reflected the competitive stage of an entertaining games where little learning had occurred. table 14. sports games: game refinement and momentum game b d p~ gr table tennis (post-2000) 54.86 96.47 0.01802898 0.077 basketball 36.38 82.01 0.01903259 0.073 soccer 2.64 22 0.01916382 0.073 badminton (current) 46.34 79.34 0.02099768 0.086 tic-tac-toe 9 9 0.10602695 0.333 western chess 35 80 0.01930085 0.074 japanese chess (shogi) 80 115 0.01654861 0.078 go 250 208 0.01152868 0.076 baghchal 242 26 0.09342471 0.075 3) application to edutainment games: educational games are played for educational purposes where the intention of the game may be to acquire a new skill (such as a language in 65 | international journal of informatics information system and computer engineering 1 (2020) 47-78 duolingo and memrise) or to check an already acquired skill (such as scrabble)2. players in the educational games grow actively throughout the game. as such, information at lower level and at higher level are taken as the game information. although scrabble is considered to be a board game, this study considered it as an educational games since its primary purpose is education. as such, the p~ value is obtained by (12) in the educational games. three educational games were considered in this study. the first one is duolingo, a language learning platform which has pre-structured course and ‘spanish for english speakers course’ is considered in this study. memrise is also language learning platform which does not have structured courses where the user can create the course according to their convenience but a pre-structured course is also available. only the language learning perspective is considered for memrise. where ‘japanese for english speakers’ is used for this study. lastly, scrabble is an educational game designed to test one’s skills where the english language version is considered. these games were chosen for their popularity and step by step progress method which assisted in the analysis of growth rate. in spite of the fact that duolingo has utilized many game components in their stage, the badges and winning streak are explicitly examined because the critical piece of effective gamification is the substance and learning material itself (boskic, n., and hu, s., 2015). also, the primary reason for players interested in duolingo involves taking in a second language, while the badges and winning streak are utilized to lift up the player motivation. with respect to the work in (huynh, d et al., 2016), the growth rate of learning and quantifying entertaining experience utilizes formulation (6). as such, the gr and p~ values for each milestone in the badge and winning streak systems are given in table 15. according to previous studies (huynh, d et al., 2016), (huynh, d., & iida, h., 2017), the gr value implies the significant of the winning streak to users’ motivation compared to the badges. the momentum in both badge and winning streak scenarios decreases to eventually become almost equal after certain milestones. this implies that duolingo is more of an educational but entertaining game as well as better for beginner level of learners but not very attractive to learners of advanced level. while winning streak element of duolingo successfully maintains uncertainty, it does not solve the problem of falling curiosity. 2 to maintain the fairness, all the three games are considered as language learning educational game. since the comparison is not between skill acquisition and testing knowledge, it does not make the difference in the results. the games are tested on an individual basis and not categorical. h iida and m n a khalid a paradigm shift from optimal...| 66 table 15. duolingo badge and winning streak: game refinement (gr) and momentum (p~) milestone badge lk → 𝒑 gr 1 6 21 0.03706 0.1166 2 16 67 0.01797 0.0597 3 29 160 0.0914 0.0337 4 61 317 0.00563 0.0246 milestone winning streak lk → 𝒑 gr 1 2.2 21 0.01631 0.0047 2 6.4 67 0.01203 0.0053 3 21.9 160 0.00838 0.0127 4 47.8 317 0.00530 0.0260 lk : average number of lessons learned; in memrise, users are required to accumulate experience point (xp) to cross a particular level. this mechanism works similar to the milestone technique in the duolingo game. similarly, the gr and the p~ are calculated where the xp refers to the required xp accumulation to reach a particular level. as evident from table xvi, p~ value is building up while the gr value is rapidly declines. from such observation, even though the user’s learning growth is maintained for a period of time and increases in another, limited game element makes it difficult to keep the user entertained. in addition, in term of catching up with the course structure, it is better for the user when the learning growth gradually increases. nevertheless, a sudden jump in p~ in memrise implies that the learning growth suddenly become too large which become difficult to catch up (i.e. level 8 to 9 and level 14 to 15). this explains the reason for the declining gr value since acquiring xp point becomes a burden instead of enjoyment which make the game lose its attractiveness for the user (especially when trying to reach level 13). thus, having diverse game elements during the game play may enhance the game progress and rejuvenate the curiosity of users. finally, scrabble may be considered as another board game but differ because of its scoring mechanism. as such, the scoring rate, the branching factor, and the game length can be observed. since those information are fundamental for applying the game progress model, the gr value can be computed utilizing the board game model, as in (9). then, the momentum (p~ value) are calculated by simulating selfplay of an open source game of scrabble called quackle version 1.0.3. the data includes the average of 100 game plays between various levels of ai players. based on the results of obtained in scrabble (fig. 4), the momentum is 67 | international journal of informatics information system and computer engineering 1 (2020) 47-78 highest when the dictionary size is the highest and the game length is the lowest. this actually aligned with the findings by (kananat, s et al., 2016) where that study claimed that scrabble is most enjoyable table 16. memrise experience point (xp): game refinement (gr) and momentum (p~) level x p lw → 𝒑 gr 1 500 11 0.69314718 2,033 2 1,000 22 0.69314718 1,437 3 2,000 44 0.69314718 1,016 4 4,000 88 0.69314718 0,719 5 8,000 177 0.69314718 0,505 6 16,000 355 0.69314718 0,356 7 32.000 711 0.69314718 0,252 8 64,000 1422 0.69314718 0,178 9 128,000 2844 0,91629073 0,126 10 320.000 7111 0,91629073 0,080 11 800,000 17777 0,91629073 0,050 12 2,000,000 44444 0,91629073 0,032 13 5,000,000 111111 0,91629073 0,020 14 12,500,000 277777 0,91629073 0,013 15 31,250,000 694444 1,16315081 0,008 xp : experience points; lw : learned words; when the dictionary size of the ai player is the highest. in contrast, smaller dictionary size and higher game length results in decrease in the value of momentum which implies that high learning growth (due to high p~ value) would occur when experienced players (corresponds to high dictionary size or maximum game information) compete within a short period of time. h iida and m n a khalid a paradigm shift from optimal...| 68 fig. 4. momentum in scrabble. it can be seen that momentum is highest when the game length is short and game information is maximum. b. comfort-in-mind looking from another perspective, the gr measure provides the whole game playing situation through a single constant value. in other word, both the game element (such as reward) and the total process are constant as well as the attractiveness of the whole process remain the same. however, if the game element along the process changes, how can the measure of attractiveness be determined in any time or step of the process? by focusing on the change of the gr value throughout the game process, its “attractive” tendency can be determined using the third derivatives of (3), resulting in (14). this process gravitates towards increase in attractiveness which represents an addictive-like situation or strong feeling of engagement, namely as the ad value given by (15). this value can be perceived as the “force” in the human mind, where the present of “inertia” makes it difficult to persist along the total process. in other word, a timely reward change is an appropriate stimulation to help users continue to overcome the difficult uphill process (potentially caused by boredom). as such, a relation between the bio-mechanical effects and psychological attractiveness can be established. for example, when we take the roller coaster, not only we can feel the gravity, but also we can feel the “attraction” in our mind. as such, the relationship between the change of gravity (phenomenon known as “weightlessness”) and the change of attractiveness (in this context, called “addiction”) is investigated. a ‘login system’ is widely used and relevant in many fields, especially in the domain of online learning games such as 69 | international journal of informatics information system and computer engineering 1 (2020) 47-78 doudizhu3, baicizhan4 and duolingo11. a login system is a typical monotonous repetitive action. as such, the system designers are expected to set some rewards to motivate the users to log in more often. from user’s point-of-view, such reward in the login system can be regarded as some form of entertainment. however, maintaining engagement in this type of game is not an easy feat, especially with the rise of free-toplay business model of games (castro, e. g., and tsuzuki, m. s., 2015) and rising popularity of elearning as an educational medium (shabihi, n et al.,2016). some game elements had been adopted to increase motivation and improve retention of the users on such system. such feats can be achieved by careful attention to the relations between the attractiveness or the loyalty of the users and the game elements (huotari, k., and hamari, j., 2012). this is especially true for an online learning system such as login system, where several tiers and different reward rules can be applied. previous study had provide evidence that “tiers system” is better than the “equal reward” because the process provides smoother decrease of user’s attractiveness throughout longer time span (months to years). based on this aspects, two popular application named “we are primary school students” and “war poetry” were adopted for this purpose. the first application regulation involves user receiving one reward per day through application login which described by two factors: the number of reward 3 http://www.auway.net/games/play/dou-dizhu.htm obtained from “continuous login days” and the number of “login days” that corresponds to the g and t, respectively. the focus is on making more gentle and smooth loss of the user’s attractiveness based on the reward rule. as such, 5 different reward rules are defined as in table 17 which represents an incremental reward system. table 17. five different reward schedules and rules. rule # period∗ reward distribution 1 365 days 1.00 2 365 days/2 0.50, 1.50 3 365 days/3 0.50, 1.00, 1.50 4 365 days/4 0.25, 0.75, 1.25, 1.75 5 365 days/5 0.25, 0.75, 1.00, 1.25, 1.75 ∗: divided into equal parts; by applying the five rules in table 17, the curves of the gr value and ad value are depicted in fig. 5. since the login application is a repeated action, losing attractiveness from the users quickly is typical and should be avoided. fig. 5 (top) shows the change of gr value over time based on the regular reward rule which rapidly declining at the beginning. this implies that the users will lose their attractiveness and feel boring in a short time. however, the decrease of gr values is the least for rule 5, which indicates a smoother “attractiveness decrease” process. although rule 1 to 3 provide better gr value in the beginning, the rapid decrease of the gr value over time would cause the inability of the user to 4 https://www.baicizhan.com 11https://www.duolingo.com/ http://www.auway.net/games/play/dou-dizhu.htm https://www.baicizhan.com/ https://www.baicizhan.com/ https://www.duolingo.com/ h iida and m n a khalid a paradigm shift from optimal...| 70 retain engagement. in addition, the change of ad values over time (see fig. 5(bottom)) also showed similar tendency of reducing smoothly, especially on rule 5 which implies gentle losses on users’ attractiveness. imagine the “weightlessness” of the roller coaster, the change of the ad values corresponds to the degree of “attractiveness-less” where it is more acceptable to be experienced by the users little by little. in other words, although “attractiveness-less” can be viewed as a negative experience, it still retains user’s sense of enjoyment, which reduces smoothly over a long period of time. as such, “attractiveness-less” that barely noticeable to the users who can make their engagement to be retained longer. another example application is the war poetry which adopted similar reward rules and schedules is given in table 18. this enables the player to obtain login rewards that grows over time, rather than remain constant every day. if the number of rewards and the login days in war poetry is regarded as the g and t, respectively, the changes of the gr and ad values (fig. 6) can be obtained. the differences of the gr values and ad values, at the beginning and the ending for each login systems is summarized in table xix. when the reward system is smoother towards losing user’s “attractiveness”, the value of ∆gr × ∆ad is lower. this implies that changes in gr and ad values are important factors to retain the engagement of the users. as such, it is important for the system designers to set different reward rules in different time stages. based on the value of ∆gr×∆ad, both the incremental reward (rule 4 and rule 5) and tiers system is the most effective in gently reducing the “attractiveness” and prolonging the engagement of the user. although the total reward is a constant, the rate of change of fig. 5. the change of gr and ad values over time: five different reward rules compared 71 | international journal of informatics information system and computer engineering 1 (2020) 47-78 table 18. different reward schedules of “war poetry” from first to the seventh day reward rule war poetry rewards point system tier system day 1 1.77 1.00 day 2 1.77 1.10 day 3 1.77 1.30 day 4 1.77 1.50 day 5 1.77 2.00 day 6 1.77 2.50 day 7 1.77 3.00 total 12.40 12.40 fig. 6. the change of gr and ad values over time: different reward rules of “war poetry” compared table 19. the analysis of gr value and add value based on login system login system grb gre adb ade ∆gr × ∆ad rule1 1.0000 0.0523 1.0000 0.0027 0.94514 rule2 0.7071 0.0523 0.7071 0.0027 0.46124 rule3 0.7071 0.0523 0.7071 0.0027 0.46124 rule4 0.5000 0.0523 0.5000 0.0027 0.22264 rule5 0.5000 0.0523 0.5000 0.0027 0.22264 point system 1.3310 0.5031 1.3310 0.1901 0.94455 tiers system 1.0000 0.5031 1.0000 0.1901 0.40244 b: begin; e: end h iida and m n a khalid a paradigm shift from optimal...| 72 the gr value had demonstrated some successful cases for the designers to set different reward rules in different stages in two example login applications, “we are primary school student” and “war poetry”. maximizing the user’s attractiveness and making them more loyal are achieved by dividing the process to many stages in order to optimize the reward rule (zuo, l., et al, 2017). looking from another perspective, the interpretation of the rate of change of the gr values and ad values can be related to the bio-mechanical processes of the roller coaster (eager, d et al., 2016). ∆gr means the interval of the attractiveness of users in the total process, which implies the range of maximum and minimum “attraction”, just like the height of the roller coaster. the ad value is similar with the feeling of “weightlessness” (or related to the force of gravity) where ∆ad implies the interval of the degree of “weightlessness” (changing rate of the gr value). making both ∆gr and ∆ad smaller would also make the “attractiveness” to decline smoothly. similarly, the tiers system which corresponds to the value of ∆gr×∆ad, also indicates smooth and progressive reduction of the gr value; thus, corresponds to the acceptability of the user. 5. general discussion applications of the gr theory have seen much success in quantifying the attractiveness and sophistication of almost all well-known competitive games such sports (fig. 7), board games (fig. 8), and video games (fig. 9). games that highly depend on luck (shorter game length, high branching factor) are situated in area greater than gr = 0.08 while games that highly depend on skill (longer game length, high branching factor) are situated in area lesser than gr = 0.07. these so called “comfortable” zone (gr ∈ [0.07,0.08]) was identified based on a shared patterns of games that found to be enjoyable and attractive to both beginner and advance level players. the zone provides a harmonic balance between the level of challenge and luck imposed on the player when playing the game. for the games that situated area gr < 0.07, it is largely dependent on skill that can be made known (or controlled) in order to influence the game outcome (i.e. the precision of hitting on the baseball bat and level of knowledge of a chess player). in contrast, games situated in area gr > 0.08 are highly dependent on luck/chance where any small external factors may not be visible (uncontrollable), but nevertheless influence the game outcome (i.e. the 73 | international journal of informatics information system and computer engineering 1 (2020) 47-78 probability of encountering rock for three consecutive time in roshambo and change of pace or direction of the shuttle in badminton). fig. 7. comparison of various sports relative to the the game refinement curves of gr ∈ [0.07,0.08]. gr = 0.07 and gr = 0.08 is the lower bound and upper bound of game “comfortable” zone, respectively. while these “comfortable” zone provide evidence the state of mind of players in the competitive environment, such zone needed to be generalized in much broader context. based on the application of the gr measure in educational standpoint, a gamified learning experience provides user with better appeal to the overall learning process where technique such as learning milestone had improved the learning experience by making learning session shorter and more engaging (see sectioniii-a). in addition, gr measure also provides indication for users to approach the educational materials in a more sensible manner (i.e. user with motivation). meanwhile, from the business standpoint, the gr measure provides insights to both the customer and the business owner with respect to their own interests and benefits (see section iii-b). for example, the gr measure for the sales promotion ‘discount’ provides insights on product popularity and business sustainability, where the zone value balances these two factors so that it becomes beneficial to both customer and the business owner. in addition, the gr measure of the starbucks customer loyalty program indicates the need for the customer to spend and the free items the starbucks willing to give to their loyal customer. similarly, the gr measure of the gamified services for hotel and airlines loyalty programs also showed that “skillful” customer (becoming loyal customer) will likely enjoy the benefits of their services while the companies systematically improve their revenues. in-depth analysis of the gr measures and its derivatives may also provide a different perspective to the state of mind of users or players. while competition in game may be related to enjoyment and entertaining state-ofmind, little was known to what perceived to happen in our mind when stimulated (such as curiosity during learning) and boredom (such as retention of engagement). application of gr measure showed that some of the gamified elements of an edutainment game may entice certain learner curiosity (such as duolingo entices beginner level learner due to its being an entertaining edutainment game) while impedes learner curiosity when lacks content diversity (see section iv-a). in addition, skillful or experienced player that compete within a short session would also encourage growth of h iida and m n a khalid a paradigm shift from optimal...| 74 learning (i.e. high curiosity). on the other hand, application of gr measure and its derivative (ad value) also provide insights on dealing with boredom. making a long and repetitive tasks engaging can be difficult, but possible when the sense of enjoyment is incrementally injected in-between tasks (see section iv-b). although repetitive tasks can be viewed as negative experience and reduce users engagement, reducing it smoothly and gradually may prolong the sense of engagement as much as possible while avoiding user rejection (or frustration). by avoiding abrupt and sudden change in the user experience, the biomechanical equivalent known as “weightlessness” is imitated from the perspective of the psychological attractiveness, where the force of acceleration (thrills or excitement of accumulating rewards) cancelled out by the inertia (rejection or frustration of repeating the repetitive tasks). such situation can be perceived as reaching the state of “attractiveness-less”, which approximately correlates with “addiction”12. in essence, the findings from the successful application of the gr measure and its derivative have provided evidence on the current paradigm shifts where the state of design for the users or players have shifted from optimizing play experience through various aspect of the game (or their environment) into state of design for the users or players that prioritized comfortable psychological state of mind in various aspects of the game (or their environment). as such, game designer not only needs to design an optimal play experience, but also a comfortable state of mind to provide a holistic approach in game development as well as other related ind ustries; thus, blurring the line between work and play (iida h., 2018). fig. 8. comparison of various board games relative to the the game refinement curves of gr ∈ [0.07,0.08]. gr = 0.07 and gr = 0.08 is the lower bound and upper bound of game “comfortable” zone, respectively. fig. 9. comparison of various video games relative to the the game refinement curves of gr ∈ [0.07,0.08]. gr = 0.07 and gr = 0.08 is the lower bound and upper bound of game “comfortable” zone, respectively. 75 | international journal of informatics information system and computer engineering 1 (2020) 47-78 6. conclusion and future prospects this paper highlights evidences of the occurring paradigm shifts that were observed from the application of the game refinement theory and its derivative in the educational and business perspective, while further explored its utility in understanding and mapping the stimulation and ultimatum of the state of human mind, through the lens of games. nevertheless, there are still plenty of open problems concerning the current research that requires further exploration. since the recent debut of alphago and alphazero, fair play becomes an important issue in game (aung, h. p. p et al., 2019). mechanically define the term “fairness” from the framework of game refinement theory provide a fascinating prospects in the future. on the other hand, with the recent development of accessible and widely available brain-computer interface technology such as muse13, exploring the correlations of the measured brain waves with respect to the game refinement theory, where actual emotional and psychological state of mind can be mapped to distinct labels for better understanding of our brain and how it work, would be an interesting direction for future research. finally, similar to the work in (browne, c., 2020). which focuses on the physical evolution of games, further exploration from the psychological footprints of game evolution through the framework of the game refinement theory (see preliminary works by (yicong, w., 2019) can be ventured in the future. acknowledgment this research is funded by a grant from the japan society for the promotion of science, within the framework of the grant-in-aid for challenging exploratory research. references agarwal, s., khalid, m. n. a., & iida, h. (2019). using curiosity model to characterize edutainment based on learner’s growth rate. in in press. agarwal. s, khalid, m. n. a. and iida h., 2019. using curiosity model to characterize edutainment based on learner’s growth rate.” eai, 10. aksakal, n. (2015). theoretical view to the approach of the edutainment. procediasocial and behavioral sciences, 186, 1232-1239. aung, h. p. p., khalid, m. n. a., & iida, h. (2019). can we save near-dying games? an approach using advantage of initiative and game refinement measures. eai. aung, h. p. p., khalid, m. n. a., & iida, h. (2019, november). towards a fairness solution of scrabble with komi system. in 2019 international conference on advanced information technologies (icait) (pp. 66-71). ieee. h iida and m n a khalid a paradigm shift from optimal...| 76 boskic, n., & hu, s. (2015, october). gamification in higher education: how we changed roles. in european conference on games based learning (vol. 1, p. 741). academic conferences international limited. browne, c. (2020). ai for ancient games. ki-künstliche intelligenz, 34(1), 89-93. castro, e. g., & tsuzuki, m. s. (2015). churn prediction in online games using players’ login records: a frequency analysis approach. ieee transactions on computational intelligence and ai in games, 7(3), 255-265. de boer, e. r., & gudmundsson, s. v. (2012). 30 years of frequent flyer programs. journal of air transport management, 24, 18-24. deterding, s., dixon, d., khaled, r., & nacke, l. (2011, september). from game design elements to gamefulness: defining" gamification". in proceedings of the 15th international academic mindtrek conference: envisioning future media environments (pp. 9-15). diah, n. m., nossal, n., zin, n. a. m., higuchi, t., & iida, h. (2014). a game informatical comparison of chess and association football (soccer). advances in computer science: an international journal, 3(4), 89-94. e. fisher. (2016). “2016’s most important customer loyalty statistic”. [online]. available: http://www.annexcloud.com/blog/2016/02/ 05/ultimatecustomer-loyalty-statistics-2016/ eager, d., pendrill, a. m., & reistad, n. (2016). beyond velocity and acceleration: jerk, snap and higher derivatives. european journal of physics, 37(6), 065008. gino, f (2018). the business case for curiosity, harvard business review (visited on 10/6/2019). [online]. available: https://hbr.org/2018/09/curiosity harsanyi, j. c., & selten, r. (1988). a general theory of equilibrium selection in games. mit press books, 1. huotari, k., & hamari, j. (2012, october). defining gamification: a service marketing perspective. in proceeding of the 16th international academic mindtrek conference, 17-22. huynh, d., & iida, h. (2017). an analysis of winning streak's effects in language course of" duolingo". huynh, d., zuo, l., & iida, h. (2016, december). analyzing gamification of “duolingo” with focus on its course structure. in international conference on games and learning alliance, 268-277. springer, cham. iida, h. (2008). fairness, judges and thrill in games. japan advanced institute of science and technology, tech. rep, 28. iida, h. (2018). where is a line between work and play?. iida, h., takahara, k., nagashima, j., kajihara, y., & hashimoto, t. (2004, september). an application of game-refinement theory to mah jong. in international conference on entertainment computing (pp. 333-338). springer, berlin, heidelberg. http://www.annexcloud.com/blog/2016/02/05/ultimate-customer-loyalty-%20statistics-2016/ http://www.annexcloud.com/blog/2016/02/05/ultimate-customer-loyalty-%20statistics-2016/ http://www.annexcloud.com/blog/2016/02/05/ultimate-customer-loyalty-%20statistics-2016/ https://hbr.org/2018/09/curiosity https://hbr.org/2018/09/curiosity 77 | international journal of informatics information system and computer engineering 1 (2020) 47-78 izzyann, f., huynh, d., xiong, s., norshakirah, a., & iida, h. (2018). comparative study: case study in analyzing gamification between mind-snacks and duolingo. in the international conference on information and communication technology and digital convergence business (icidb-2018). j. huizinga. (2014). homo ludens ils 86. routledge. kananat, s., terrillon, j. c., & iida, h. (2016, december). gamification and scrabble. in international conference on games and learning alliance, 405-414. springer, cham. kumar, k. (2016). an analysis on brand loyalty: a case study on starbucks. maor, e. (2019). the pythagorean theorem: a 4,000-year history. princeton university press. neumann, j. v. (1928). zur theorie der gesellschaftsspiele. mathematische annalen, 100(1), 295-320. panumate, c., iida, h., & takahashi, r. (2017). an analysis of sports using game refinement measure. odisha journal of social science, 4. panumate, c., iida, h., & terrillon, j. c. (2016, november). a game informatical analysis of roshambo. in 2016 conference on technologies and applications of artificial intelligence (taai), 116-123. panumate, c., xiong, s., & iida, h. (2015, july). an approach to quantifying pokemon's entertainment impact with focus on battle. in 2015 3rd international conference on applied computing and information technology/2nd international conference on computational science and intelligence, 60-66. punyawee, a., panumate, c., & iida, h. (2016). finding comfortable settings of snake game using game refinement measurement. in advances in computer science and ubiquitous computing, 66-73. springer, singapore. ramadhan, a., iida, h., & maulidevi, n. u. (2015). game refinement theory and multiplayer games: case study using uno. shabihi, n., taghiyareh, f., & abdoli, m. h. (2016, september). analyzing the effect of game-elements in e-learning environments through mbti-based personalization. in 2016 8th international symposium on telecommunications (ist), 612-618 sutiono, a. p., purwarianti, a., & iida, h. (2014, july). a mathematical model of game refinement. in international conference on intelligent technologies for interactive entertainment, 148-151. springer, cham. takeuchi, j., ramadan, r., & iida, h. (2014). game refinement theory and its application to volleyball. information processing society of japan, 2014, 1-6. terrill, b. 2008. my coverage of lobby of the social gaming summit. (visited on 10/12/2019). [online]. available: http://www.bretterrill.com/ 2008/06/mycoverage-of-lobby-of-social-gaming.html http://www.bretterrill.com/2008/06/my-coverage-of-lobby-of-social-gaming.html http://www.bretterrill.com/2008/06/my-coverage-of-lobby-of-social-gaming.html http://www.bretterrill.com/2008/06/my-coverage-of-lobby-of-social-gaming.html h iida and m n a khalid a paradigm shift from optimal...| 78 xin, o. w., zuo, l., iida, h., & aziz, n. (2017, november). gamification effect of loyalty program and its assessment using game refinement measure: case study on starbucks. in international conference on computational science and technology (pp. 161-171). springer, singapore. xiong, s., & iida, h. (2014, november). attractiveness of real time strategy games. in the 2014 2nd international conference on systems and informatics (icsai 2014) (pp. 271-276). ieee. xiong, s., mao, x., li, w., & iida, h. (2017). finding comfortable settings of mafia game using game refinement measurement. 情報処理学会研究報告 (web), 2017(gi-38). xiong, s., zuo, l., & iida, h. (2014). quantifying engagement of electronic sports game. advances in social and behavioral sciences, 5, 37-42. xiong, s., zuo, l., & iida, h. (2017, september). possible interpretations for game refinement measure. in international conference on entertainment computing (pp. 322-334). springer, cham. xiong, s., zuo, l., chiewvanichakorn, r., & iida, h. (2014, october). quantifying engagement of various games. in the 19th game programming workshop 2014. information processing society of japan. yicong, w., aung, h. p. p., khalid, m. n. a., & iida, h. (2019, november). evolution of games towards the discovery of noble uncertainty. in 2019 international conference on advanced information technologies (icait) (pp. 72-77). ieee. zuo, l., & iida, h. (2017, september). an analysis of sales promotion ‘discount’using game refinement measurement. in international conference on entertainment computing (pp. 487-491). springer, cham. zuo, l., xiong, s., & iida, h. (2017, november). an analysis of hotel loyalty program with a focus on the tiers and points system. in 2017 4th international conference on systems and informatics (icsai) (pp. 507-512). ieee. zuo, l., xiong, s., wang, z., & iida, h. (2018, june). an analysis of gamification effect of frequent-flyer program. in international conference on e-learning and games (pp. 53-60). springer, cham.