关键词: 协同过滤, 信任, 推荐, 移动通信

Abstract: The traditional collaborative filtering technology carries on the recommendation mainly based on user-item dataset, and cannot efficiently use the contextual information of user communication, thereby the recommended accuracy is further constrained. Aiming at the shortcomings of traditional collaborative filtering recommendation algorithm, in this paper, it fuses communication contextual information into the collaborative filtering algorithm, and introduces three types trust including communication trust, similarity trust and transmission trust, and a trust-based collaborative filtering recommendation model is also proposed. Experiments on the public dataset demonstrate that the recommendation algorithm outperforms the traditional collaborative algorithm.

Key words: collaborative filtering, trust, recommendation, mobile communication