A CHAT-BOT FOR PROVIDING RECOMMENDATIONS FOR WATCHING VIDEOS BASED ON MATRIX FACTORIZATION MODELS
DOI:
https://doi.org/10.32689/maup.it.2024.1.3Keywords:
recommendation system, chatbot, matrix factorization, sigular value decomposition method, machine learning, field-aware factorization machineAbstract
The article examines the main models of predicting user reactions in recommendation systems based on matrix factorization methods. The choice of the matrix factorization model is justified and the approaches to ensuring the flexibility of interaction between the recommendation system and the user through the use of a chatbot implemented in web applications are considered. The purpose of the article is to study the effectiveness of using a chatbot in providing personalized recommendations for viewing video content on the basis of a matrix factorization model. Research methods. General methods of developing web applications and intelligent chatbots are used, methods of matrix factorization using SVD sigular value decomposition method, machine learning methods, natural language processing and recognition methods, and recommendation system optimization methods based on assessment of forecast accuracy, satisfaction level of communication with the chatbot. The scientific novelty of the study consists in the identification of methods and approaches aimed at improving users' receipt of personalized recommendations for watching video content in accordance with their interests and preferences by using a chatbot and a model for predicting user reactions based on matrix factorization methods. Conclusions. The accumulation of large volumes of digital video information in various formats requires the improvement of mechanisms for providing recommendations and increasing the accuracy of providing predictions regarding user preferences. The research of the matrix factorization models MF, the factorization machine FM, and the field-aware factorization machine FFM made it possible to establish that the model of the field-aware factorization machine FFM had the best indicators of forecast accuracy: MAE=0,86, MSE=1,65, RMSE=1,28. To ensure the flexibility of user interaction with the recommendation system developed on the basis of the FFM model, the expediency of its integration with a chatbot implemented in the web application was found. The research of the quality of the created natural language processing model showed a high accuracy of recognizing the user's intentions when communicating with the chatbot - 99.17%. Detection of the level of user satisfaction with communication with the chatbot and received recommendations made it possible to establish that user satisfaction was 86.6%. Which indicates a high level of assessment of the effectiveness of user interaction with the chatbot and the high accuracy of the system in terms of predicting users' intentions to watch videos.
References
Болюбаш Н. М. Інтелектуальний аналіз даних. Миколаїв: Вид-во ЧДУ ім. П. Могили, 2023. 320 с. URL: https://dspace.chmnu.edu.ua/jspui/handle/123456789/1461
Мелешко Є. В., Семенов С. Г., Хох В. Д. Дослідження методів побудови рекомендаційних систем в мережі Інтернет. Системи управління, навігації та зв'язку. Вип. 1(47). 2018. С. 131–136. URL: https://doi.org/10.26906/SUNZ.2018.1.131
Мелешко Є. В., Хох В. Д., Босько В. В. Дослідження матричних факторизаційних моделей рекомендаційних систем. Системи управління, навігації та зв’язку. Вип. 6(58), 2019. С. 58–62. URL: https://doi.org/10.26906/SUNZ.2019.6.058
Abbas M., Riaz M. U., Rauf A., Khan M. T., Khalid S. Context-aware Youtube recommender system. In International Conference on Information and Communication Technologies ( ICICT). Karachi: IEEE, 2017. P. 161–164. URL: https://doi.org/10.1109/ICICT.2017.8320183
Attalariq M., Baizal Z.K.A. Chatbot-Based Book Recommender System Using Singular Value Decomposition. Journal of Information System Research. Vol. 4, No 4. 2023. P. 1293–1301. URL: https://doi.org/10.47065/josh.v4i4.3817
Chin W. S., Zhuang Y., Juan Y. C., Lin C. J. A Learning-Rate Schedule for Stochastic Gradient Methods to Matrix Factorization. In Advances in Knowledge Discovery and Data Mining. 19th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2015). Lecture Notes in Computer Science. Vol. 9077. 2015. Springer Cham. P. 442–445. URL: https://doi.org/10.1007/978-3-319-18038-0_35
Dalton J., Ajayi V. Main R. Vote Goat: Conversational Movie Recommendation. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval. Ann Arbor MI USA: ACM, 2018. P. 1285–1288. URL: https://doi.org/10.1145/3209978.3210168
Duchi J., Hazan E., Singer Y. Adaptive Subgradient Methods for Online Learning and Stochastic Optimization. Journal of Machine Learning Research. Vol 12. 2011. P. 2121–2159. URL: https://jmlr.org/papers/volume12/duchi11a/duchi11a.pdf
Fajari A. N., Baizal A. Chatbot-based Culinary Tourism Recommender System Using Named Entity Recognition. Journal Imiah Penelitian dan Pembelajaran Informatika. Vol. 7, No. 4. 2022. P. 1131–1138. URL: https://doi.org/10.29100/jipi.v7i4.3210
Falk K. Practical recommender systems. Shelter Island, NY: Manning, 2019. 432 p.
Fayyaz Z., Ebrahimian M., Nawara D., Ibrahim A., Kashef R. Recommendation Systems: Algorithms, Challenges, Metrics, and Business Opportunities. Applied Sciences. Vol. 10(21). 2020. URL: https://doi.org/10.3390/app10217748
Gomez-Uribe C.A., Hunt N. The netflix recommender system: algorithms, business value, and innovation. ACM Transactions on Management Information Systems. Vol. 6, No 4. 2016. P. 1–19. URL: https://dl.acm.org/doi/10.1145/2843948
Hong F.X., Zheng X.L., Chen C.C. Latent space regularization for recommender systems. Information Science. Vol. 360. 2016. P. 202–216. URL: https://doi.org/10.1016/j.ins.2016.04.042
Jannach D. Evaluating conversational recommender systems: A landscape of research. Artificial Intelligence Reviev. Vol. 56, No. 3. 2023. P. 2365–2400. URL: https://doi.org/10.1007/s10462-022-10229-x
Jayalakshmi S., Ganesh N., C¨ep R., Senthil Murugan J. (2022). Movie recommender systems: concepts, methods, challenges, and future directions. Sensors. Vol. 22(13). URL: https://doi.org/10.3390/s22134904
Juan Y., Zhuang Y., Chin W.S. Field-aware Factorization Machines for CTR Prediction. In Proceedings of the 10th ACM Conference on Recommender Systems. 2016. P. 43–50. URL: https://doi.org/10.1145/2959100.2959134
Ma S., Zha Z., Wu F. Knowing user better: jointly predicting click-through and playtime for micro-video. In IEEE International Conference on Multimedia and Expo ( ICME). 2019. P. 472–477. URL: https://doi.org/10.1109/ICME.2019.00088
ML.NET Documentation. URL: https://learn.microsoft.com/en-us/dotnet/machine-learning
MovieLens 25M Dataset. URL: https://grouplens.org/datasets/movielens/25m
Narducci F., Gemmis M.D., Lops P., Semeraro G. Improving the User Experience with a Conversational Recommender System. In AI*IA Advances in Artificial Intelligence. 2018. P. 528–538. URL: https://doi.org/10.1007/978-3-030-03840-3_39
Nugraha M., Baizal Z.K.A., Richasdy D. Chatbot-Based Movie Recommender System Using POS Tagging. Building of Informatics, Technology and Science (BITS). 2022. Vol. 4, No. 2. P. 624–630. URL: https://doi.org/10.47065/bits.v4i2.1908
Padti P. G., Hegde K., Kumar P. Hybrid Movie Recommender System. International Journal of Research in Engineering, Science and Management. Vol. 4, no. 7, 2021. P. 311–314.
Pujahari A., Sisodia D. S. Model-Based Collaborative Filtering for Recommender Systems: An Empirical Survey. In 2020 First International Conference on Power, Control and Computing Technologies (ICPC2T). Raipur, India: IEEE, 2020. P. 443–447. URL: https://doi.org/10.1109/ICPC2T48082.2020.9071454
Singh A., Ramasubramanian K, Shivam S. Building an Enterprise Chatbot: Work with Protected Enterprise Data Using Open Source Frameworks. Berkeley, CA: Apress. 2019. 385 p. URL: https://doi.org/10.1007/978-1-4842-5034-1
Sun J., Zhang A., Pan J., Flores A. Field-matrixed Factorization Machines for Recommender Systems. In WWW’21: Proceedings of the Web Conference. 2021. P. 2828–2837. URL: https://doi.org/10.1145/3442381.3449930
Taddy M. Stochastic Gradient Descent. In Business Data Science: Combining Machine Learning and Economics to Optimize, Automate, and Accelerate Business Decisions. New York: McGraw-Hill. 2019. P. 303–307.
Theosaksomo D., Widyantoro D. H. Conversational Recommender System Chatbot Based on Functional Requirement. In IEEE 13th International Conference on Telecommunication Systems, Services, and Applications (TSSA). 2019. P. 154–159. URL: https://doi.org/10.1109/TSSA48701.2019.8985467
Welcome to pytorch-accelerated’s documentation. URL: https://pytorch-accelerated.readthedocs.io/en/latest
Zhang Y. An Introduction to Matrix factorization and Factorization Machines in Recommendation System, and Beyond. 2022. URL: https://doi.org/10.48550/arXiv.2203.11026
Zhang Z., Lui Y., Zhang Zh. Field-Aware Matrix Factorization for Recommender Systems. IEEE Access. Vol. 6. 2018. P. 45690–45698. URL: https://doi.org/10.1109/ACCESS.2017.2787741
Zhao X., Li X., Liao L., Song D., Cheung W.K. Crafting a time-aware point-of-interest recommendation via pairwise interaction tensor factorization. In Knowledge Science, Engineering and Management. KSEM. 2015. Lecture Notes in Computer Science. Vol. 9403. Cham, Switzerland: Springer. P. 458–470. URL: https://doi.org/10.1007/978-3-319-25159-2_41
