Abstract: Providing high quality recommendations for users is a significant topic in e-commerce environment, however, it suffers from data sparse problem. To address the problem, this paper proposes a collaborative filtering recommendation algorithm based on trust propagation. The algorithm proposes TSR weight combining trust, similarity and relationship to replace the similarity in traditional collaborative filtering algorithm in order to find neighbours. TSRCF algorithm solves the data sparse problem and helps users get high quality recommendations in the information overload environment. Experimental results based on Epinions data sets and FilmTrust data sets demonstrate that the algorithm performs better than the traditional filtering algorithm in terms of accuracy.

Key words: trust propagation, sparsity, collaborative filtering, weight Trust, Similarity, Relationship（TSR）

摘要： 电子商务环境下为用户提供高效的推荐是一个非常有意义的课题，然而稀疏性问题严重影响了推荐系统的推荐质量。为了有效解决这个问题，提出了一种基于信任传播的TSRCF协同过滤算法，在信任传播的基础上，提出了信任度，相似度，关系度的混合权重TSR，取代了传统的协同过滤算法的相似度，作为寻找邻居用户的标准。TSRCF算法在一定程度上缓解了稀疏性问题，帮助用户在信息过载的情境下得到高质量的推荐。在Epinions数据集和FilmTrust数据集上的仿真实验也验证了TSRCF算法比传统CF算法有更高的推荐精确度。

关键词: 信任传播, 稀疏性, 协同过滤, 信任度、相似度、关系度的混合权重