622 https://internationalpubls.com Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) Final Assignment Title Recommendation System Using Collaborative Filtering Method for Undergraduate Students of Informatics Engineering Anang Pramono1*, Devi Fitrianah2 12Postgraduate Program in Informatics Engineering, Computer Science Department, BINUS Graduate Program Master of Computer Science, Bina Nusantara University, Jl. K. H. Syahdan No. 9, Kemanggisan, Palmerah, Jakarta 11480, Indonesia Email: 1*anang.pramono@binus.ac.id, 2devi.fitrianah@binus.ac.id Article History: Received: 16-10-2024 Revised: 27-11-2024 Accepted: 12-12-2024 Abstract: The final project title selection process is a crucial stage for undergraduate students of Informatics Engineering. The difficulty in determining a title that is relevant to their interests and abilities is often an obstacle. This research aims to develop a final project title recommendation system that can assist students in finding suitable research topics. This system utilizes a collaborative filtering approach to analyze historical data on course grades and student preferences. The results show that the item-based collaborative filtering method produces more accurate and relevant recommendations compared to the user-based collaborative filtering method. The RMSE value of the item-based collaborative filtering method (0.4985) is lower than the RMSE value of the user-based collaborative filtering method (0.9759). While the accuracy value of the item-based collaborative filtering method (0.5903) is higher than the user-based collaborative filtering method (0.0153).This system has the potential to optimize the final project title selection process, accelerate study time, and improve the quality of student research. Keywords: Recommendation system, final project, collaborative filtering INTRODUCTION The process of preparing the final project is one of the important stages in the academic journey of students, especially at the undergraduate level. Choosing the right final project title, which is relevant to students' abilities and preferences, has a significant influence on the quality of the final result and the success of the study. A suitable title will motivate students to explore the topic in depth, increase productivity, and ultimately produce quality scientific work. However, in practice, there are still many students who have difficulty in determining the right final project title. Some of the factors that cause this include the large number of titles. The more title options available, the more difficult it is for students to make a decision, the second is due to a lack of understanding of the field of study. Not all students have a deep understanding of all fields of study in their study program. While the third factor is the limited time factor. The process of searching and selecting titles often takes a long time, making it difficult for students to balance it with other academic activities. Difficulty in choosing a final project title often results in delays in study completion. Students who struggle to find interesting and relevant topics tend to experience difficulties in the research and writing Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 623 https://internationalpubls.com process, resulting in the time needed to complete the final project being longer than it should be. Not only does this result in delays in graduation, but it can also increase students' stress levels and frustration. To overcome these problems, a system is needed that can help students find final project titles that match their abilities and preferences. One approach that can be used is the collaborative filtering method. This method has been widely used in recommendation systems in various domains, such as e-commerce and streaming services. The basic principle of this method is to recommend items (in this case, final project titles) to users based on the similarity of preferences with other users or items that have been previously interacted with. LITERATURE REVIEW Recommendation System A recommendation system is a system that automatically suggests items (products, services, information) to a user based on that user's preferences, behavior, or characteristics. The main goal of a recommender system is to help users find items that are relevant and interesting to them amidst the many options available. Types of Recommendation Systems • Content-based filtering: This system recommends items based on the similarity between an item that the user already likes and other items that have similar attributes. For example, if a user likes sci-fi movies, the system will recommend other sci-fi movies. • Collaborative filtering: This system recommends items based on the similarity of preferences between users. There are two main approaches in collaborative filtering: o User-based collaborative filtering: Searches for other users who have similar preferences to the target user, and recommends items favored by these similar users. o Item-based collaborative filtering: Searches for items that are similar to items the user already likes, and recommends similar items. • Hybrid filtering: A combination of content-based and collaborative filtering. This approach seeks to combine the strengths of both methods to produce more accurate recommendations. Recommendation System Application Recommendation systems have been widely applied in various fields, such as: • E-commerce: Recommending products to customers based on purchase history, browsing behavior, or demographic characteristics. • Movies and music: Compile playlists or movie recommendations based on user preferences. • News and articles: Displays news or articles relevant to the user's interests. • Social media: Suggests new friends or content that may be of interest to the user. • Education system: Recommend learning materials that match students' level of understanding and interest. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 624 https://internationalpubls.com Collaborative Filtering Method Basic Principles of Collaborative Filtering Collaborative filtering works by analyzing the interaction patterns between users and items. The system attempts to find similarities between users or items to provide relevant recommendations. Comparison of User-based and Item-based Collaborative Filtering • User-based: o Pros: Able to capture unique user preferences. o Disadvantages: Prone to sparsity problem (lack of user interaction data), more complex computation, and difficult to handle new users. • Item-based: o Pros: More computationally efficient, more resistant to sparsity problems. o Disadvantages: Lacks the ability to capture unique user preferences. Commonly Used Algorithms • User-based: o Neighborhood-based: Calculates the similarity between users based on the ratings given on the same item. o Model-based: Using modeling techniques such as matrix factorization to predict unknown ratings. • Item-based: o Item-based nearest neighbor: Searches for items that have similar characteristics to items that the user already likes. Cosine Similarity • Definition: Cosine similarity is a metric that measures the similarity between two vectors. In the context of collaborative filtering, these vectors represent user preferences or item characteristics. • Calculation: Performed by calculating the cosine of the angle between two vectors. The cosine similarity value ranges from -1 to 1, where a value of 1 indicates perfect similarity. • Applicability: Used to calculate the similarity between users or items in collaborative filtering. Performance Measurement • RMSE (Root Mean Squared Error): Measures the average difference between predicted and actual values. A small RMSE value indicates a good prediction model. • Accuracy: The proportion of correct predictions to total predictions. • Precision: The proportion of correct positive predictions to all positive predictions. • Recall: The proportion of correct positive predictions to all true positive examples. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 625 https://internationalpubls.com • F1-score: Harmonic mean of precision and recall, providing a balance between precision and recall. RESEARCH METHOD In this research, we will develop a final project title recommendation system that utilizes the collaborative filtering method. This system aims to help undergraduate students of Informatics Engineering in choosing a final project title that is relevant to their interests and abilities. Research Design This study uses an experimental research design. Thus, we will build and evaluate the performance of the developed recommendation system. Object of Research The research object in this study is undergraduate students of Informatics Engineering in the Informatics Engineering Study Program, Universitas 17 Agustus 1945 Surabaya, 2020 - 2024 period. Research Variables • Independent Variable: o Collaborative filtering methods (user-based, item-based) o Similarity calculation algorithm (cosine similarity) o Amount of training data • Dependent Variable: o Recommendation accuracy (RMSE, MAE) o Accuracy, Precision, Recall, F1-score Data Collection Methods The data used in this study are: • Student data: Consists of student identity data, course grades, and research interests. This data will be used to build student profiles. • Final project title data: Consists of a list of available final project titles. Data related to student grades are grouped into 4 research groups. With this grouping, courses that have relevance to the research group will be made into one research group. The results of the average score from each group will be compared, the higher score will be a reference to the system so that it will produce the main recommendation for students on the given title. Table 1. Grouping of courses in the research group No. Research Group Course Content 1 Information System 1. Object-Oriented Programming 2. Data Structure 3. Database Management Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 626 https://internationalpubls.com 4. Data Structure 5. Software Testing 6. Data Structures and Algorithms 2 Image Processing and Artificial Intelligence 1. Image Processing 2. Computational Intelligence 3. Data Minning 4. Deep Learning 5. Graphs and Automata 6. Discrete Mathematics and Logic 7. Computational Math 8. Natural language processing 9. Inference Statistics 3 Hardware, Networking and IoT 1. Digital System 2. Computer Architecture and Organization 3. Computer Network System 4. Emerging Technology 5. IoT 6. Robotics 7. Jarkom Administration 8. Cyber Security System 4 Multimedia and Mobile Applications 1. Human and Computer Interaction 2. Computer Graphics 3. Web Programming 4. Game Development 5. Mobile App Development Algorithm and Implementation • Data pre-processing: The acquired data will be cleaned, converted into a suitable format, and normalized. • Model building: A collaborative filtering model will be built based on the processed data. • Evaluation: The built model will be evaluated using predefined evaluation metrics (RMSE, MAE, precision, recall, F1-score). Research Flow 1. Data collection: Collected student data, final project titles, and student preferences. 2. Data pre-processing: Cleaning and converting data into a suitable format. 3. Model building: Build a collaborative filtering model using the selected algorithm. 4. Evaluation: Evaluate the performance of the model using testing data. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 627 https://internationalpubls.com 5. Analysis of results: Analyze the evaluation results and compare the performance of various methods. Figure 1. Research flow The application of the collaborative filtering method in this final project title recommendation system can be described in Figure 1. Figure 2. Collaborative Filtering method approach Scope and Limitation • Scope: This research focuses on developing a final project title recommendation system using the collaborative filtering method for undergraduate students of Informatics Engineering. Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 628 https://internationalpubls.com • Limitation: This research only considers historical data of course grades and student interests as factors determining recommendations. External factors such as current research trends, student activities in campus activities are not considered in this research. RESULT AND DISCUSSION Experiment Results In this research, some data was utilized. The research also obtained the results as presented below. • Data Statistics Table 2. Statistical Results of Data No. Item data Description 1 Judiciary Student Data Period 2020-2024 2 Student GPA Data Period 2020-2024 3 Research Group Data 4 Groups • Comparison of Method Performance Table 3. RMSE testing of item-based collaborative filtering method Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 629 https://internationalpubls.com Table 4. Testing the user-based collaborative filtering method Figure 3. Visualization graph of method performance comparison Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 630 https://internationalpubls.com Table 5. Comparison of testing item and user-based collaborative filtering methods Table 6. average value of Accuracy, Precision, Recall and F1 Score testing CF Method Accuracy Precision Recall` F1 Score Item Based 0.5903 0.6012 0.6184 0.6255 User Based 0.0153 0.0261 0.023 0.0179 Discussion Based on the experimental results, the item-based collaborative filtering method performs better than the user-based method in terms of prediction accuracy, as indicated by the lower RMSE value. This is likely due to the sparsity problem in the data used, so the item-based method, which is more resistant to sparsity problems, gives better results. CONCLUSION This research successfully developed a collaborative filtering-based final project title recommendation system for undergraduate students of Informatics Engineering. This system is designed to assist students in choosing a final project title that is relevant to their interests and abilities. The evaluation results show that the item-based collaborative filtering method provides better performance than the Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 631 https://internationalpubls.com user-based collaborative filtering method with a value of 0.4985 or lower than the user-based collaborative filtering method. Implications The development of this recommendation system has several implications: • Time efficiency: Students can save time in finding a suitable final project title. • Increased relevance: Recommended final project titles have higher relevance to students' interests and abilities. • Increased motivation: Students will be more motivated to work on their final project if the topic chosen matches their interests. • Improving the quality of the final project: By choosing a relevant title, students are expected to produce a better quality final project. Suggestions for Future Research This research still has some limitations that can be the direction of development for further research, among others: • Feature development: The system can be developed by adding new features such as recommendations based on trending topics or recommendations based on students' programming skills. • Use of more complex algorithms: The use of hybrid algorithms that combine content-based filtering and collaborative filtering to improve recommendation accuracy can be considered. • Improved data quality: The quality of the data used greatly affects the performance of the system. Therefore, efforts should be made to improve the quality of the data used. • More comprehensive evaluation: Evaluations can be conducted by involving more users and using different scenarios. The final project title recommendation system developed in this research is a good first step in helping undergraduate students of Informatics Engineering in choosing research topics. With further development, this system is expected to be a useful tool in improving the quality of higher education. REFERENCES [1] Alabduljabbar, R., Almazrou, H., & Aldawod, A. (2023). Context-Aware News Recommendation Sistem: Incorporating Contextual Information and Collaborative Filtering Techniques. International Journal of Computational Intelligence Sistems, 16(1). https://doi.org/10.1007/s44196-023-00315-5 [2] Aljunid, M. F., & Huchaiah, M. D. (2020). Multi-model deep learning approach for Collaborative Filtering recommendation sistem. CAAI Transactions on Intelligence Technology, 5(4), 276–282. https://doi.org/10.1049/trit.2020.0031 [3] Boström, P. & Filipsson, M., 2017, Comparison of User Based and Item Based Collaborative Filtering Recommendation Services, KTH R. Inst. Technol, pp. 1-9. [4] Chen, R., Hua, Q., Chang, Y. S., Wang, B., Zhang, L., & Kong, X. (2018). A survey of Collaborative Filtering -based recommender sistems: from traditional methods to hybrid methods based on social networks. IEEE Access, 6(February), 64301–64320. https://doi.org/10.1109/ACCESS.2018.2877208 Communications on Applied Nonlinear Analysis ISSN: 1074-133X Vol 32 No. 5s (2025) 632 https://internationalpubls.com [5] Feixiang, X. (2024). Intelligent Personalized Recommendation Method Based on Optimized Collaborative Filtering Algorithm in Primary and Secondary Education Resource Sistem. IEEE Access, 12(January), 28860–28872. https://doi.org/10.1109/ACCESS.2024.3365549 [6] Geetha, G., Safa, M., Fancy, C., & Saranya, D. (2018). A Hybrid Approach using Collaborative Filtering and Content based Filtering for Recommender Sistem. Journal of Physics: Conference Series, 1000(1). https://doi.org/10.1088/1742-6596/1000/1/012101 [7] I Gede Iwan Sudipa, Suyono, Jefri Junifer, D. (2017). Sistem Pendukung Keputusan. In Angewandte Chemie International Edition, 6(11), 951–952. http://repo.iain-tulungagung.ac.id/5510/5/BAB 2.pdf [8] Kirubahari, R., & Miruna Joe Amali, S. (2021). A Hybrid Deep Collaborative Filtering Approach for Recommender Sistems. Research Square. https://doi.org/10.21203/rs.3.rs-651522/v1 [9] Kumar, P., Gupta, M. K., Rao, C. R. S., Bhavsingh, M., & Srilakshmi, M. (2023). A Comparative Analysis of Collaborative Filtering Similarity Measurements for Recommendation Sistems. International Journal on Recent and Innovation Trends in Computing and Communication, 11(December 2022), 184–192. https://doi.org/10.17762/ijritcc.v11i3s.6180 [10] Kumar, H., Singh, M. K., Gupta, M. P., & Madaan, J. (2020). Moving towards smart cities: Solutions that lead to the Smart City Transformation Framework. Technological Forecasting and Social Change, 153(October 2017), 1–16 [11] Ngatno. (2015). Metodelogi Penelitian Bisnis. Semarang: Lembaga dan Penjaminan Mutu Pendidikan Unverstas Diponegoro. [12] Pagare, R., & Shinde, A. (2012). A Study of Recommender Sistem Techniques. International Journal of Computer Applications, 47(16), 1–4. https://doi.org/10.5120/7269-0078 [13] Pangesti, W. E., Suryadithia, R., Faisal, M., Wahid, B. A., & Putra, A. S. (2021). Collaborative Filtering Based Recommender Sistems For Marketplace Applications. International Journal of Educational Research & Social Sciences (IJERSC), 2(5), 1201–1209. https://ijersc.org [14] Patoulia, A. A., Kiourtis, A., Mavrogiorgou, A., & Kyriazis, D. (2023). A Comparative Study of Collaborative Filtering in Product Recommendation. Emerging Science Journal, 7(1), 1–15. https://doi.org/10.28991/ESJ-2023- 07-01-01 [15] Rajalakshmi, S., & Santha, K. R. (2023). Hybrid Recommender Sistem Using Systolic Tree for Pattern Mining. Computer Sistems Science and Engineering, 44(2), 1251–1262. https://doi.org/10.32604/csse.2023.024036 [16] Ricci, F., Rokach, L., Shapira, B., Kantor, P. B., & Ricci, F. (2011). Recommender Sistems Handbook. In Recommender Sistems Handbook. https://doi.org/10.1007/978-0-387-85820-3 [17] Sun, S., Yu, Z., Cui, X., & Zhou, Y. (2023). Collaborative Filtering with Graph Neural Networks: A Survey. IEEE Transactions on Big Data, 10(1), 1-17. [18] Tian, L., & Liu, F. (2023). Design of Music Heritage Database Sistem Based on Collaborative Filtering Algorithm. Frontiers in Artificial Intelligence and Applications, 373, 438–443. https://doi.org/10.3233/FAIA230839 [19] Upadhyaya, P. (2023). Study of Mathematical Model for User-based Collaborative Filtering and Item. January. https://www.researchgate.net/publication/366902172 [20] Zhang, S., Yao, L., Sun, A., & Tay, Y. (2019). Deep learning based recommender sistem: A survey and new perspectives. ACM Computing Surveys, 52(1). https://doi.org/10.1145/3285029 [21] Zhou, Z., Zhang, L., & Yang, N. (2023). Contrastive Collaborative Filtering for Cold-Start Item Recommendation. In ACM Web Conference 2023 - Proceedings of the World Wide Web Conference, WWW 2023 (Vol. 1, Issue 1). Association for Computing Machinery. https://doi.org/10.1145/3543507.3583286 [22] Wang, C., Boughanem, M., & O'Connor, T. (2022). Collaborative Filtering for Context-Aware Recommendations. ACM Transactions on Intelligent Sistems and Technology, 13(2), 1-23. [23] Wang, X., He, X., Wang, X., & Nie, Z. (2022). Collaborative Filtering with Heterogeneous Information. IEEE Transactions on Knowledge and Data Engineering, 34(11), 3200-3214. [24] Wu, X., Zhou, T., Li, W., & Sun, Y. (2021). Collaborative Filtering with Privacy-Preserving Techniques. ACM Transactions on Information and Sistem Security, 24(4), 1-26.