Final Year Report Template Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 262 https://internationalpubls.com Predicting Disasters: A Machine Learning Approach Dr.Rayavarapu Veeranjaneyulu1, Dr. V. Sumathi2, Dr. C. Sushama3, Savanam Chandra Sekhar4, P. Neelima5, Dr. M. Sunil Kumar6 1Associate Professor, Department of CSE, PACE Institute of Technology & Sciences, Ongole,Prakasam,AP,India. veerupace@gmail.com 2Associate Professor, Department of Mathematics, Sri Sai Ram Engineering College, Chennai,TN,India. sumathi.math@sairam.edu.in 3Associate professor, Department of CSE, School of Computing, Mohan Babu University, (erstwhile Sree Vidyanikethan Engineering College), Tirupathi, AP, India. Susyma83@gmail.com 4Associate Professor, Department of BBA, KL Business School, Koneru Lakshmaiah Education Foundation, KL University, Vaddeswaram ,Guntur Dt., AP, India. savanam.sekhar@gmail.com 5Assistant professor, Department of CSE, School of engineering and technology, Spmvv, tirupati, Andhra Pradesh. neelima.pannem@gmail.com 6Professor , Department of Computer Science and Engineering, School of Computing, Mohan Babu University(erstwhile Sree Vidyanikethan Engineering College),Tirupati, AP, India. sunilmalchi1@gmail.com Article History: Received: 15-08-2024 Revised: 28-09-2024 Accepted: 15-10-2024 Abstract: Trends in river water levels that could cause floods are an intriguing and applicable subject of study. They set up shop to lessen the financial and social impact of floods. Water level variations can be predicted using support Vector, a class of machine learning methods that can also be used to detect the likelihood of flooding. The two algorithms use comparable hydrological and flood resource variables in the process of predicting floods, such as the prediction amount, River inflow, seasonal flow, flood frequency, and other pertinent flood prediction variables, because the water level is the most important factor in hydrological research. Machine learning techniques are helpful for predicting floods because they have the ability to classify and regress data from several sources into groups of flood and non-flood. Anyone may use the Machine Learning and Deep Learning features of artificial intelligence thanks to Watson Studio. Keywords: forecasting, Disaster Management, machine learning, deep learning. I. INTRODUCTION India consistently ranks as the most flood-prone country year. Cities' low-lying areas are more vulnerable to flooding. There are a number of important variables that have led to the increase of water logging, including surface runoff, relative height, and insufficient passage of the water to drainage. Flood forecasting is especially important there. Several Indian states have been hit by flooding in the past year. These include Assam, Bihar, Goa, Odisha, Pune, Maharashtra, Tamil Nadu, Karnataka, Kerala, and Gujarat. Chennai received 1049 mm more precipitation than average during November 2015, the wettest month of the year. Rainfall of 1088 millimeters in November was the highest since records began in 1918. The Kanchipuram district receives an average of 64 centimeters of precipitation between October and December. The highest rainfall total was recorded there at 181.5 cm, which is 183% greater than normal rainfall. However, 146 cm of rain was measured in the mailto:sunilmalchi1@gmail.com Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 263 https://internationalpubls.com Tiruvallur district. There has been a lot of effort put into developing systems for foreseeing floods, but not all of them provide particularly reliable estimates. Machine learning offers a plethora of solutions to improve the accuracy of problem predictions. In this study, we propose estimating the flash flood as a means of protecting flood-prone areas. The strategy is geared on developing a functional model of an ML algorithm. Improved short-term forecasting in urban areas is achieved by factoring in the risk of flooding[6][7][27]. Gauge sensors and remote sensing devices are only two of the many tools available for tracking floods. However, there are limitations to remote sensing technologies, such as the length of time between satellite visits, and the limited spatial resolution of gauge sensors. Recovering from a data loss might take anywhere from a few minutes to a few hours. Due to its reliance on water level sensors and precipitation forecasts, existing flood warning analysis is unable to offer near-real-time and automated flood monitoring analysis. As a result, visual sensing devices have been created, capable of collecting vast amounts of data in a targeted area. There are countless applications that make substantial use of the information included in both moving and still photographs. With the advancement of surveillance technology, more people are considering installing visual monitoring and surveillance systems, especially in disaster-stricken areas.[8][25][26] II. LITERATURE SURVEY J.Akshaya and P.L.K. Priyadarshini [1] The recent flooding in several parts of southern India has caused significant loss of life and property. Resuming normal life after a devastating natural disaster like a flood might take a long time. Drones are only one example of the technology used during disasters to speed up relief efforts and reduce property loss. Remote sensing and aerial picture analysis automatically need a large number of algorithms. Drones with high-quality cameras and sensitive sensors are replacing traditional aerial photography in many applications. This research suggests a hybrid method for determining whether or not an area in an aerial photograph has been impacted by flooding. Support Vector Machine (SVM) and k-means clustering, when used together, were able to properly categorize around 92% of flooded photos, demonstrating their efficacy in detecting flooded areas. Altering the SVM's kernel functions is how performance is analyzed. The results demonstrate that using quadratic SVM cuts down on both prediction and training times[11][12]. A.B.Ranjit and P.V.Durga [2] One of the most pressing and difficult issues in hydrology is flood forecasting (FF). Forecasting and warning for flooding is widely recognized as the single most effective non-structural measure for mitigating flood losses. A reliable flood warning system should alert residents with the time to prepare. The purpose of an accurate flood forecast is to give authorities and citizens as much warning as possible before a flood strikes. The development of hydrological models, the expansion of analytical expertise, and the enhancement of data collecting via satellite observations have all contributed to an increase in the accuracy of forecasts. In this work, we examine several facets of flood forecasting, including the models employed, data collection methods, visualization strategies, and alerts[9][28][29]. F. A. Ruslan, K. Haroon, A. M. Samad and R. Adnan, [2017] Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 264 https://internationalpubls.com Many states in Malaysia, especially those on the east coast, are vulnerable to devastating floods at the end of the year because of Monsoon rain. Numerous people saw significant financial and property losses. In light of this, an alert system predicated on a reliable model for predicting flood water levels is required to give advance warning to the impacted area and its residents. Models for predicting floods were developed using the Multiple-Input, Single-Output (MISO) Auto regressive with Exogenous Input (ARX) and the Multiple-Input, Single-Output (MISO) Auto regressive Moving Average with Exogenous Input (ARMAX) architectures, and their predictive capacities were evaluated. The parametric models were developed with the help of the MATLAB System Identification toolkit. Information for this case study was gathered at five different locations along the Pahang River in Temerloh, Pahang: four upstream sites and one downstream station. The data utilized in this analysis were collected and compiled by the Malaysian Ministry of Drainage and Irrigation. In terms of Best Fit value and rise values, the simulation results showed that the ARMAX structure-designed flood prediction model performed better than the ARX structure-designed model[10][19][20]. [3]In many parts of the world, floods are among the most devastating natural disasters that may strike. Particularly in the Philippines, this was a huge problem because of the potential for material loss, infrastructural destruction, and even human casualties. Current systems strictly adhere to problems to prevent disastrous flood disasters. In this research, we use real-time data from monitoring devices to create a model that can forecast when and where floods will occur. Based on data from sensors in a real-time monitoring system, the system makes accurate flood level predictions in advance. The prediction model was created using a multi-layered artificial neural network and MATLAB. The network's goodness-of-fit was excellent across all four datasets (0.99889 for the training set, 0.99362 for the test set, 0.99764 for the validation set, and 0.99795 for the full dataset)[13][14]. III. EXISTING MODEL When rivers, creeks, and other waterways transport an excessive amount of water to a region that lacks sufficient drainage, flooding results. In the last 20 years, floods have harmed over 2.3 billion people, causing countless deaths, destroying over 92 billion head of cattle annually, damaging seven million hectares of land, and costing over $5 trillion worldwide in just the last five years[15][16]. Limited data availability and the definition of events can contribute to data-related difficulties. Difficulties in drought and flood models arise from a lack of data, which makes it impossible to describe the entire diversity of extremes. When doing extreme value analysis, the decision on which event description to utilize is critical. Observations of stream flow are used in both statistical and hydrological models for purposes of calibration and evaluation. However, it's possible that these records don't exist, exist only temporarily, or have holes in them. It's also possible that data aren't readily available to the public because of issues like measurement mistakes or in homogeneities. It's also possible that we're missing spatial data and details about the variables that affect stream flow. We explain these types and how they influence flood and drought prediction models in the following paragraphs[17][18]. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 265 https://internationalpubls.com No stream flow observations: This difficulty is especially felt on the African continent due to the fact that observed stream flow records are only available at the precise areas along a stream where a gauge has been installed and station density may be low[21][23]. Data access: Due to data licensing limitations, rigorous access controls, or the time needed to make these datasets readily useable at the global scale, publicly available stream flow datasets are still uncommon. Storage in decentralized databases, generally maintained by regional rather than national authorities and often only accessible in the local language, further impedes data access[24] IV. PROPOSED SYSTEM In proposed system, we implement a Machine Learning algorithm for getting insights from the complex patterns in the data. This technique is computationally inexpensive because of its simple architecture. And here we are using XG Boost Algorithm for great Data Access and Accuracy. In order to manage environmental and water resource systems, as well as to set insurance rates and evacuate individuals from flood-prone areas, accurate flood forecasts are essential. Floods and droughts are two of the most common natural calamities in India, both of which may be traced back to the unpredictable nature of the monsoon. Highest death reported due to flood& drought. Need some solution to give prior indication of flood & drought. Data Set The data set consists of 630 records and features are Date District, State Mar-May, Jun-Sep, Oct- Dec, of type string and Numeric the dataset is derived from the Kaggle.com and google websites. By using these dataset we can Predict the floods these dataset will be in the csv format. By using dataset we can predict the future floods. Data Training Our model is trained with XG Boost and other machine learning techniques using a pre-processed training dataset. The viewer then views the categorized results. Measures to compute The dataset is trained using a variety of different algorithms, including SVM, LR, KNN, and MLP. The calculating variable. 1) Precision: Precision refers to the rate at which a positive prediction is corroborated by data. True Positive (TP) and False Positive (FP) both stand for the same thing. 2) Recall: Are call is the ratio of observed positive values that is correctly predicted. Formula: TP/(TP+FN) 3) F1Score: F1score takes the (1) and (2)for the calculation. It gives the average weight. Formula: F1score=2(Precision*Recall)(Recall+Precision) Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 266 https://internationalpubls.com 4) Sensitivity: Sensitivity is also called a Recall. TP becomes higher when the sensitivity value is higher. Prediction Results: Class 1: Floods Found Class 2:Floods not found If the given water millimeter is higher than 2400 it will be 1 or if it is lesser than 2400 then it is 2. Figure 1: Prediction Result The figure shows the steps taken before separating a huge data set into two distinct groups. Both a training and a test dataset. We use the Logistic Regression, Support Vector Machine, K-nearest, and XG boost classification algorithms in the training data set, and the Multilayer Perceptron (MLP) model in the testing data set. Then results will get with higher accuracy in binary values either flood may happen or Floods may not happen. This is the way we will predict the results. User Register The user needs to register. About-Project In this application, we have successfully created an application which takes in and at a set and predicts the Floods happen or Floods not happen in the Particular areas. Login A registered user can login using the valid credentials to the website to use a application. Upload Dataset The user has to upload an dataset inform at of csv files which needs to be tested. Prediction The results of our model are displayed as either Floods happen or not. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 267 https://internationalpubls.com View Results View the results that floods happen or not. Logout Once the prediction is over, the user can logout of the application Figure 2 -Architecture In this architecture first we are uploading dataset after uploading raw data we have to preprocess it in 2ratios 80:20 . 80 percent is training dataset and 20 percent is testing. And classification we are using is XGBoost Algorithm. And we are getting the results Floods Happen or Floods Not Happen. V. IMPLEMENTATION METHODS Logistic Regression: Logistic regression is a statistical model. Here we use Probability technique of 1 event out of 2 alternatives. Here in logistic regression here we use only binary values either 0 or 1 to model the probability of certain class or event taking place. In our prediction we use logistic regression it’s prediction tie is0.9288. it will tell only either Floods may happen or Floods may not happen. In threshold it will show as S curve. The Categorical responses are only two outcomes. KNN Algorithm: KNN algorithm is an supervised learning algorithm. Here it can use for both classification and regression technique to assign weights. It Takes the nearest values If K=3 if distance is close to one classifier then it will move to 1 classifier. Here It calculates Distance and Weights. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 268 https://internationalpubls.com Figure 3.-KNNAlgorithm Here we can see that in this fig the blue point is a new data point. Category a and category b are neighbors. This algorithm calculating nearest distance new data point is assigned to category because the distance is very less comparing to category b. Support Vector Machine: In Support vector machine we can analyze the data by using either Classification technique or Regression model Here We have 2 categories to train the data set if new data comes it can easily categorize under a particular group. It is called as non-Binary Linear Classifier. In Support vector machine we train a large data with all the features. And then we will give them to one strange data by using hyper planes Support vector machine will find the best route or best boundary decision. In support vector machine we use ML classifier to predict the floods but its accuracy is 0.9666. So with is algorithm our measures are not giving best accuracy. Figure 4 : Support Vector System In this Figure we can see the best hyper plane boundary and X1 and X2 are two independent variables. By using KNN Classifier and by using previous data we can find the best margin. This hyper plane margin is Soft Margin. XG Boost algorithm: In our Prediction modeling we are using XG Boost algorithm. Because it’s regularized boosting and prevents over fitting. It can handle missing values automatically. This is a parallel processing method with the added benefit of iterative cross-verification. It lets you set your own optimization goals and do tree pruning, as well as identify the ideal number of iterations. Causes trees to grow deeper but more optimally Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 269 https://internationalpubls.com FrameWork: Here we use the Frame Work called PyCharm by using Prediction modeling techniques we install the package called XG Boost and also interfaces NumPy files Sickit_learn. The reason it doesn't share Scikit_learn's APIs is that it was built independently. Features and labels are stored in a D Matrix Structure. Using & all parameters supplied via a dictionary, we can quickly and easily generate it from a NumPy array. Then we'll refer to the data as "trained" in order to make accurate predictions. XG Boost Hyper parameters: Here we use booster technique Gbtree orgb linear objectives of hyper parameters is multi: SoftMax, and multi: Softprob and the learning rate is Adjusting Weights and also here we are calculating Max depth of the trees and Minimum Child weights of the trees. And the data set splits into 2 phases training set and testing sets 80% using dataset as training set and 20% as testing dataset parameters are Class _Weight, max_depth, min_depth. We are measuring the power of prediction in various algorithm here comparing to all algorithms XG Boost giving the highest accuracy because it has over fitting feature and boosting, bagging techniques there is an optimization and improvement in this picture when we compare to logistic regression and XG boost 0.9662 is the XGboost accuracy. And the training time in seconds is 24 seconds. This is the prediction power of XGBoost algorithm. Because of this it provides cache awareness. In the Figure We are measuring the power of prediction in various algorithm here comparing to all algorithms XGBoost giving the highest accuracy because it has over fitting feature and boosting, bagging techniques there is an optimization and improvement in this picture when we compare to logistic regression and XG boost0.9662 is the XGboost accuracy. And the training time in seconds is 24 seconds. This is the prediction power of XGBoost algorithm. Because of this it providing cache awareness. Figure 5: - XG-Boost Algorithm VI. EXPERIMENT RESULT Analysis of various Classification algorithms to predict the floods: Table 1- experimental result Parameter Logistic Regression Support Vector Machine K-Nearest Algorithm XG- Boostalgorithm Precision 0.66 0.87 0.86 0.988 Recall 0.87 0.89 0.88 1 F1-Score 0.82 0.90 0.89 0.98 Sensitivity 0.87 0 0.87 0.99 Accuracy 90.3 93.2 95.85 97.4 Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 270 https://internationalpubls.com In the above table we can see that various analysis of classification Algorithms for Flood Prediction Here we are using parameters to calculate the accuracy to predict the floods. In Logistic Regression 0.66 is the precision in SVM 0.87, and K-N 0.86, but in XG boost algorithm has highest precision 0.988 Compare to all the algorithms XGboost giving better accuracy to predict the floods. Recall is a parameter to find the ratio values compare to the other entire algorithms XG boost giving more accuracy. The accuracy is 1. F1 score is used to calculate average weight of recall and precision after averaging accuracy is 0.98. So as per analysis we can see the measures XG boost giving better accuracy within 30sec we can find the values for the data. Table 2: INPUT PARAMETERS S.no Mar-May Jun-Sep 10days_june Rainfall happen(0) 1 386.2 2122.8 274.8666667 649.9 0 2 275.7 2403.4 130.3 256.4 1 3 336.3 2343 186.2 308.9 0 4 339.4 2398.2 366.0666667 862.5 0 5 378.5 1881.5 283.4 586.9 0 6 230 1943.1 138.3 254.1 0 7 328 2737.8 256.9666667 669.5 1 8 283.7 2023.6 197.5333333 450 0 9 628.3 1940.4 234.9 231.5 0 10 296.7 1886.5 226.6666667 531.2 0 11 303.1 2167 232.0333333 541.6 0 12 240.4 1851.7 234.5666667 580.8 0 13 767 1104.3 154.7666667 218.7 0 14 346.8 2025 212.2666667 389.8 0 15 283.7 2318.2 321.4333333 876.6 0 16 202.3 2928.4 240.8333333 642.5 0 Floods Happen (1) Floods not 10 REPORTS S.NO User Name Algorith Used Percentage Floods Happen Floods Not Happen 1 Vineeta XGBoost Algorithm 0.9999 NO yes 2 Vineela Logistic Regression 0.8787 No yes 3 Shabana Decision Classifier 0.88686 yes No 4 Deepthi KNN Algorithm 0.87898 Yes No 5 Vinay XGBoost algorithm 0.9898 yes No 6 Akshaya XGBoost algorithm 0.9898 yes No 7 Ahalya XGBoost algorithm 0.989899 No Yes Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 271 https://internationalpubls.com 8 Amulya XGBoost Algorithm 0.99999 No Yes 9 Lucky XGBoost algorithm 0.9898 yes No 10 Vandana XGBoost algorithm 0.9898 yes No 11 Gourav Decision Classifier 0.88686 yes No 12 Sahithi Decision Classifier 0.88686 yes No 13 Harika LogisticmRegression 0.8787 No yes CONCLUSION Floods are Natural Disasters we cannot predict. It is impossible to Prevent the Floods with100% Accuracy but the fact is we can prevent it 0988% using XG boost prevention model. By using this we can prevent some of economical losses before floods occurrence. Every year more than 200000 people are losing their lives because of floods we can prevent it by alarming method of flood prediction. Here we used many algorithms like Logical regression, K-Nearest algorithm, Support vector machine, XGBoost algorithm. In this K-nearest algorithm accuracy LIMITATIONS To function as intended, every system must be constrained in some way. While developing this project, we ran into a few issues and constraints, such as the following: This project requires constant access to the internet; without it, we can't do anything. FUTUREWORK This can be utilized in future to classify the types of different algorithm and we are planning to invent mobile application alarming system so that rural areas can also access this prediction technique before the flood occurrence. References [1] J.AkashiandP.Priya,"AnIOTwithMachineLearningbasedFloodPredictionsystem,"2017InternationalConferenceon ICinData Science, [2] A. Rani and PV. Durga, Techniques of Flood Forecasting and Their Applications," 2018 International Conference on Mother TeresaCollege of Engineering,2018, [3] Cruzan, M Binloa "Predicting the floods by suing artificial neural networks by using water level moisture sensors TENCON 2018-2018 IEEE Region 10 Conference,2018 [4] G. Kaveri and A. Balakrishna, "Predicting the Floods by using Cloud Environment," 2019 IEEE Canadian Conference of Electrical and Computer Engineering(CCECE), [5] J.A. Ozella and A. A. Hern, Predicting Floods using Neural networks and Support vector machine: A Conceptual Framework,"2019 IEEE15thInternationalColloquiumonSignalProcessing& Its Applications (CSPA). [6] Quay Zhang., Nthabi Nidetch., Rakia, "Predicting the floods by using image processing as a real time basis. 2018 IEEE 5th International Conference on Smart Instrumentation, Measurement and Applications ICSIMA2018. [7] Rio Netsuke and Hirose, "Flood Detection in urban areas Area –A Case Study In2015 Jose Flood, Indonesia," 2018. [8] .ShijinLI. KaiakoMA. ,ZhouJIN.,andYulingZH U,"AnewMet hod flood prediction using boosting and SVM algorithms. [9] Ananthanatarajan, V., M. Sunil Kumar, and V. Tamizhazhagan. "Forecasting of wind power using lstm recurrent neural network." Journal of Green Engineering 10 (2020). [10] Natarajan, V. Anantha, et al. "Segmentation of nuclei in histopathology images using fully convolutional deep neural architecture." 2020 International Conference on computing and information technology (ICCIT-1441). IEEE, 2020. [11] Natarajan, V. Anantha, and Poojitha Karatampati. "Survey on renewable energy forecasting using different techniques." 2019 2nd International Conference on Power and Embedded Drive Control (ICPEDC). IEEE, 2019. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 1s (2025) 272 https://internationalpubls.com [12] Natarajan, V. Anantha, and N. Sandhya Kumari. "Wind power forecasting using parallel random forest algorithm." Soft Computing for Problem Solving: SocProS 2018, Volume 1. Springer Singapore, 2020. [13] Natarajan, V. Anantha, K. Saiharsha, and M. Santhosh Kumar. "Box-office revenue estimation for Telugu movie industry using predictive analytic techniques." International Journal of Recent Technology and Engineering (2019): 2277-3878. [14] Sreedhar, B., BE, M. S., & Kumar, M. S. (2020, October). A comparative study of melanoma skin cancer detection in traditional and current image processing techniques. In 2020 Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud)(I-SMAC) (pp. 654-658). IEEE. [15] Ganesh, D., Kumar, T. P., & Kumar, M. S. (2021). Optimised Levenshtein centroid cross‐layer defence for multi‐hop cognitive radio networks. IET Communications, 15(2), 245-256. [16] Balaji, K., P. Sai Kiran, and M. Sunil Kumar. "Resource aware virtual machine placement in IaaS cloud using bio- inspired firefly algorithm." Journal of Green Engineering 10 (2020): 9315-9327. [17] Kumar, M. Sunil, and K. Jyothi Prakash. "Internet of things: IETF protocols, algorithms and applications." Int. J. Innov. Technol. Explor. Eng 8.11 (2019): 2853-2857. [18] AnanthaNatarajan, V., Kumar, M. S., & Tamizhazhagan, V. (2020). Forecasting of Wind Power using LSTM Recurrent Neural Network. Journal of Green Engineering, 10. [19] Sangamithra, B., Neelima, P., & Kumar, M. S. (2017, April). A memetic algorithm for multi objective vehicle routing problem with time windows. In 2017 IEEE International Conference on Electrical, Instrumentation and Communication Engineering (ICEICE) (pp. 1-8). IEEE. [20] Balaji, K., P. Sai Kiran, and M. Sunil Kumar. "Resource aware virtual machine placement in IaaS cloud using bio- inspired firefly algorithm." Journal of Green Engineering 10 (2020): 9315-9327. [21] Ganesh, Davanam, Thummala Pavan Kumar, and Malchi Sunil Kumar. "Optimised Levenshtein centroid cross‐layer defence for multi‐hop cognitive radio networks." IET Communications 15.2 (2021): 245-256. [22] Kumar, M. S., & Harshitha, D. (2019). Process innovation methods on business process reengineering. Int. J. Innov. Technol. Explor. Eng. [23] Sreedhar, B., BE, M. S., & Kumar, M. S. (2020, October). A comparative study of melanoma skin cancer detection in traditional and current image processing techniques. In 2020 Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud)(I-SMAC) (pp. 654-658). IEEE. [24] R.Leelavathy , M. Sowmya Vani , M. Lavanya. (2022). Effectiveness Of Weight Bearing Exercise On Menopausal Problems Among Menopausal Women. Journal of Pharmaceutical Negative Results, 3301–3314. [25] M. Lavanya, M. Kalpana Devi , M. Sowmya Vani. (2022). Deep Learning For Identification Of Plantnutrient Deficiencies. Journal of Pharmaceutical Negative Results, 3284–3291. [26] A.Jyothi Babu"Self secured model for cloud based IOT systems", Measurement: Sensors, Volume 24, December 2022,https://doi.org/10.1016/j.measen.2022.100490 [27] Sushama et.al "Global Identification Passport: A Unique Cloud based Passport Model,Proceedings - 5th International Conference on Smart Systems and Inventive Technology, ICSSIT 2023, 2023, pp. 801–805 [28] Sushama et.al, "Impact of COVID-19 pandemic and the diagnosis of the virus in the human body",World Journal of Engineeringthis link is disabled, 2022, 19(5), pp. 652–657 [29] Sushama et.al, Automated extraction of non-functional requirements from text files: A supervised learning approach", Handbook of Intelligent Computing and Optimization for Sustainable Development, 2022, pp. 149–170