Social Recommendation with Biased Regularization

Authors

Xiang Hu, State Key Laboratory of Network and Switching, Beijing University of Posts and Telecommunications, Beijing, 100876, China\\ School of Control and Computer Engineering, North China Electric Power University, Beijing, 100876, ChinaFollow
Wendong Wang, State Key Laboratory of Network and Switching, Beijing University of Posts and Telecommunications, Beijing, 100876, ChinaFollow
Xiangyang Gong, State Key Laboratory of Network and Switching, Beijing University of Posts and Telecommunications, Beijing, 100876, ChinaFollow
Bai Wang, School of Computer Science, Beijing University of Posts and Telecommunications, Beijing, 100876, ChinaFollow
Xirong Que, State Key Laboratory of Network and Switching, Beijing University of Posts and Telecommunications, Beijing, 100876, ChinaFollow
Hongke Xia, Computer School, Beijing Information Science & Technology University, Beijing, 100876, ChinaFollow

Author Country (or Countries)

China

Abstract

Although recommendation systems are the most important methods for resolving the ”information overload” problem, majority of them are beset by their inherent flaws.With the recent emergence of online social networks, the increasing social information has offered opportunities to relieve these flaws. In this paper, a new matrix factorization based social recommendation method is proposed, in which social relations and the rating habit are integrated into the objective function via appending additional penalty term and bias term to classic probabilistic matrix factorization model. In order to involve more social information into traditional recommendation system, the proposed method adopt the social similarity rather than interest similarity to measure the closeness degree between users. Experiment shows that our method has got better performances than homologous methods.

Recommended Citation

Hu, Xiang; Wang, Wendong; Gong, Xiangyang; Wang, Bai; Que, Xirong; and Xia, Hongke (2015) "Social Recommendation with Biased Regularization," Applied Mathematics & Information Sciences: Vol. 09: Iss. 5, Article 48.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol09/iss5/48