Leveraging friend and group information to improve social recommender system

摘要

In recent years, we have witnessed a flourish of social commerce services. Online users can easily share their experiences on products or services with friends. Social recom-mender systems are employed to tailor right products for user needs. However, existing recommendation methods try to consider the social information to improve the recommendation performance while they do not differ the impact of different social information and do not have deep analysis on social information. In this paper, we propose a social recommendation framework to leverage the friend and group information to extend the traditional BPR model from different perspectives. Through a detailed experiment on LAST.FM data set, we find that the proposed methods are effective in improving the recommendation accuracy and we also have a good understanding for the impact of friend and group information on recommendation performance.