Trust-based Collaborative Filtering Recommendation in E-commerce

摘要：

Collaborative filtering is the dominant techniques used by todays E-commerce recommender systems, but the sparsity problem of the user-item matrix is one of the main limitations of collaborative filtering. To deal with the sparsity problem of collaborative filtering, this paper proposes a trust-based collaborative filtering algorithm. This method uses usersratings for items to calculate the direct trust between users, then based on trust inferences produces a trust matrix to find the nearest neighbors and make recommendations for a given user. Compared with traditional collaborative filtering, the proposed method can provide additional information to help alleviate sparsity. The experimental results show that the trust-based collaborative filtering algorithm can significantly improve recommendation performance.

关键词： recommender systems collaborative filtering trust inferences sparsity

作者: Rui Miao Lu Liu Haitao Xiong

作者单位: School of Economics and Management, Beihang University, Beijing, 100191, China

会议类型: 国际会议

会议名称: 第八届武汉电子商务国际会议(The Eighth Wuhan International Conference on E-Business)

会议地点: 武汉

会议语种:英文

页码: 190-195

在线出版日期: 2009-05-30（万方平台首次上网日期，不代表论文的发表时间）