پیش‌بینی امتیازات کاربران در سیستم‌های پیشنهاددهنده با درنظرگرفتن پویایی علایق کاربران و تغییرات ویژگی‌های اقلام
محورهای موضوعی : مهندسی برق و کامپیوتر
1 - گروه مهندسي كامپيوتر ، دانشگاه آزاد اسلامی واحد كاشمر، ایران
کلید واژه: سیستم‌های پیشنهاددهنده, پویایی علایق, پویایی اقلام, پیش‌بینی امتیازات, تجزیه نامنفی ماتریس.,
چکیده مقاله :
سیستم‌های پیشنهاددهنده برای استخراج اطلاعات مفید از حجم انبوهی از دادههای پیچیده به کاربران کمک کرده و استفاده از آنها در سال‌های اخیر مورد توجه‌ی چشم‌گیری قرار گرفته است. در عمل معمولا ًعلایق کاربران و ویژگی‌های اقلام در این سیستم‌ها در طول زمان تغییر می‌کنند و بنابراین تطبیق سیستم‌های پیشنهاددهنده با این نوع تغییرات ضروری بوده و به ارائه‌ی پیشنهاداتی دقیق‌تر به کاربران کمک می‌کند. با این وجود، اغلب سیستم‌های پیشنهاددهنده‌ی پویا، فقط مبتنی بر پویایی علایق کاربران در طول زمان هستند و تغییرات ویژگی‌های اقلام را در نظر نمی‌گیرند. در این مقاله، مدلی مبتنی بر تجزیه‌ی نامنفی ماتریس برای پیش‌بینی امتیازات کاربران به اقلام در سیستم‌های پیشنهاددهنده ارائه می‌شود که از هر دوی پویایی علایق کاربران و تغییرات ویژگی‌های اقلام در طول زمان استفاده می‌کند. در مدل پیشنهادی به منظور کاهش مشکل خلوتی داده‌ها، علاوه بر امتیازات کاربران از اطلاعات مربوط به اعتماد بین کاربران نیز استفاده می‌شود. نتایج ارزیابی بر روی مجموعه داده‌ی Epinions نشان می‌دهد که مدل پیشنهادی نسبت به روش‌های مورد مقایسه از دقت بهتری برخوردار می‌باشد.
Recommender systems help users to extract useful information from a large volume of complex data, and their use has received significant attention in recent years. In practice, the interests of users and the characteristics of items in these systems change over time, and therefore, adapting recommender systems to these types of changes is necessary and helps to provide more accurate recommendations to users. However, most temporal recommender systems are only based on the dynamics of users' preferences over time and do not consider changes in item characteristics.
In this paper, we propose a non-negative matrix factorization-based recommender system that uses both dynamics of users' interests and the changes in item characteristics over time in predicting users' ratings of items. In the proposed model, in order to reduce the data sparsity problem, in addition to users' ratings, trust between users is also used. The evaluation results on the Epinions dataset show that the proposed model is more accurate than the compared methods.
[1] H. Tahmasbi, M. Jalali, and H. Shakeri, "TSCMF: temporal and social collective matrix factorization model for recommender systems," J. Intell. Inf. Syst., vol. 56, no. 1, pp. 169-187, Feb. 2021.
[2] Y. Koren, "Collaborative filtering with temporal dynamics," Commun. ACM, vol. 53, no. 4, pp. 89-97, Apr. 2010.
[3] I. Rabiu, N. Salim, A. Da'u, A. Osman, and M. Nasser, "Exploiting dynamic changes from latent features to improve recommendation using temporal matrix factorization," Egypt. Informatics J., vol. 22, no. 3, pp. 285-294, Sept. 2021.
[4] J. Zhang and X. Lu, "A multi-trans matrix factorization model with improved time weight in temporal recommender systems," IEEE Access, vol. 8, pp. 2408-2416, 2020.
[5] Z. Jin, Y. Zhang, W. Mu, W. Wang, and H. Jin, "Leveraging the dynamic changes from items to improve recommendation," in Proc. 37th Int. Conf. on Conceptual Modeling, pp. 507-520, Xi'an, China, 22-25 Oct. 2018.
[6] T. Wu, Y. Feng, J. Sang, B. Qiang, and Y. Wang, "A novel recommendation algorithm incorporating temporal dynamics, reviews and item correlation," IEICE Trans. Inf. Syst., vol. E101-D, no. 8, pp. 2027-2034, 2018.
[7] I. Rabiu, N. Salim, A. Da'u, and A. Osman, "Recommender system based on temporal models: a systematic review," Appl. Sci., vol. 10, no. 7, pp. 1-27, Apr. 2020.
[8] I. Rabiu, N. Salim, A. Da'u, and M. Nasser, "Modeling sentimental bias and temporal dynamics for adaptive deep recommendation system," Expert Syst. Appl., vol. 191, no. C, Article ID: 116262, Apr. 2022.
[9] D. Rafailidis, P. Kefalas, and Y. Manolopoulos, "Preference dynamics with multimodal user-item interactions in social media recommendation," Expert Syst. Appl., vol. 74, no. C, pp. 11-18, May 2017.
[10] Y. Y. Lo, W. Liao, C. S. Chang, and Y. C. Lee, "Temporal matrix factorization for tracking concept drift in individual user preferences," IEEE Trans. Comput. Soc. Syst., vol. 5, no. 1, pp. 156-168, Mar. 2018.
[11] C. Wangwatcharakul and S. Wongthanavasu, "A novel temporal recommender system based on multiple transitions in user preference drift and topic review evolution," Expert Syst. Appl., vol. 241, no. C, Article ID: 115626, Apr. 2021.
[12 ح. طهماسبي، م. جلالي و ح. شاكري، "یک سامانه پیشنهاددهنده اجتماعی مبتنی بر تجزیه ماتریس با درنظرگرفتن پویایی علایق کاربران، "پردازش علائم و داده‌ها، سال 18، شماره 1، صص. 28-13، خرداد 1400.
[13] S. Sheibani, H. Shakeri, and R. Sheibani, "Four-dimensional trust propagation model for improving the accuracy of recommender systems," J. Supercomput., vol. 79, pp. 16793-16820, 2023.
[14] C. Zhang, K. Wang, H. Yu, J. Sun, and E. P. Lim, "Latent factor transition for dynamic collaborative filtering," in Proc. of the SIAM Int. Conf. on Data Mining, pp. 452-460, Philadelfia, PA, USA24-26 Apr. 2014.
[15] C. Zhang, Improving Recommender Systems with Rich Side Information, Simon Fraser University, 2015.
[16] Z. Chen and S. Wang, "A review on matrix completion for recommender systems," Knowl. Inf. Syst., vol. 64, no. 1, pp. 1-34, Jan. 2022.
[17] B. Ju, Y. Qian, M. Ye, R. Ni, and C. Zhu, "Using dynamic multi-task non-negative matrix factorization to detect the evolution of user preferences in collaborative filtering," PLoS One, vol. 10, no. 8, Article ID: 0135090, 2015.
[18] H. Bao, Q. Li, S. S. Liao, S. Song, and H. Gao, "A new temporal and social PMF-based method to predict users' interests in micro-blogging," Decis. Support Syst., vol. 55, no. 3, pp. 698-709, Jun. 2013.
[19] A. Y. Aravkin, K. R. Varshney, and L. Yang, "Dynamic matrix factorization with social influence," in Proc. IEEE Int. Workshop on Machine Learning for Signal Processing, 6 pp., Vietri sul Mare, Italy, 13-16 Sept. 2016.
[20] D. Rafailidis and A. Nanopoulos, "Modeling users' preference dynamics and side information in recommender systems," IEEE Trans. Syst. Man, Cybern. Syst., vol. 46, no. 6, pp. 782-792, Jun. 2016.
[21] L. Xiong, X. Chen, T. K. Huang, J. Schneider, and J. G. Carbonell, "Temporal collaborative filtering with bayesian probabilistic tensor factorization," in Proc. of the SIAM Int. Conf. on Data Mining, pp. 211-222, Colombus, OH, USA, 29 Apr.-1 May 2010.
[22] J. Zhao, S. Yang, H. Huo, Q. Sun, and X. Geng, "TBTF: an effective time-varying bias tensor factorization algorithm for recommender system," Appl. Intell., vol. 51, no. 7, pp. 4933-4944, 2021.
[23] H. Tahmasbi, M. Jalali, and H. Shakeri, "Modeling user preference dynamics with coupled tensor factorization for social media recommendation," J. Ambient Intell. Humaniz. Comput., vol. 12, no. 10, pp. 9693-9712, 2021.
[24] S. Yu, Z. Zhou, B. Chen, and L. Cao, "Generalized temporal similarity-based nonnegative tensor decomposition for modeling transition matrix of dynamic collaborative filtering," Inf. Sci. (Ny), vol. 632, pp. 340-357, Jun. 2023.
[25] C. Tong, J. Qi, Y. Lian, J. Niu, and J. J. P. C. Rodrigues, "TimeTrustSVD: a collaborative filtering model integrating time, trust and rating information," Futur. Gener. Comput. Syst., vol. 93, pp. 933-941, Apr. 2019.
[26] D. Lee and H. S. Seung, "Algorithms for non-negative matrix factorization," in Proc. of the 13th Int. Conf. on Neural Information Processing Systems, pp. 535-541, Denver, CO, USA, 1-1 Jan. 2000.
[27] D. D. Lee and H. S. Seung, "Learning the parts of objects by non-negative matrix factorization," Nature, vol. 401, pp. 788-791, 1999.
[28] H. W. Kuhn and A. W. Tucker, "Nonlinear programming," in Proc. of the 2nd Berkeley Symp. on Mathematical Statistics and Probability, vol. 1951, pp. 481-492, 1951.
[29] J. Tang, Epinions Dataset, [Online]. Available: http://www.cse.msu.edu/~tangjili/trust.html. [Accessed: 11-Feb-2022].
[30] H. Ma, T. C. Zhou, M. R. Lyu, and I. King, "Improving recommender systems by incorporating contextual information," ACM Trans. Inf. Syst., vol. 29, no. 2, Article ID: 9, 23 pp., Apr. 2011.