基于用户信任度和社会相似度的协作过滤算法  

Collaborative filtering algorithm based on user trust and social similarity

作　　者：杨海月[1] 朱玉婷[1] 施化吉[1] 徐慧 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013 [2]大全集团,江苏扬中212211

出　　处：《电子技术应用》2016年第1期100-103,共4页Application of Electronic Technique

基　　金：江苏省科技支撑计划(BE2011156)

摘　　要：个性化推荐算法是解决社交网络中信息过载问题的一种有效方法,已成为社交网络中的研究热点。协作过滤算法是被广泛应用的个性化推荐算法,但由于未考虑社交网络的一些重要社交信息及数据稀疏问题,故其在解决社交网络的推荐问题时推荐效果不佳。为此,提出一个基于用户信任度和社会相似度的协作过滤算法。首先根据用户-项目矩阵计算用户相似度,然后通过社交网络计算用户信任度和社会相似度并将三者融合,最后根据融合后的值形成最近邻集,并据此产生推荐结果。经实验分析,文中提出的算法较其他算法在解决社交网络的推荐问题时有更高的推荐精度。Personalized recommendation algorithm as an effective method to solve the information overload problem has become a re-search hotspot in social networks.Without considering some important social information of social networks and the data sparsity,the collaborative filtering algorithm which is a widely used personalized recommendation algorithm has poor recommendation effect for recommendation issues of social networks.Therefore,this paper proposes a collaborative filtering algorithm based on user trust and social similarity.Firstly,the algorithm calculates user similarity according to the user-item matrix and calculates user trust as well as social similarity through constructed user network.Next,the user similarity,user trust and social similarity will be merged to form a comprehensive value,which is used to produce neighbors.Accordingly,recommendations are produced.The experimental re-sults show that the proposed algorithm has higher recommendation accuracy than other algorithms in solving the recommendation is-sues of social networks.

关 键 词：协作过滤 数据稀疏 用户信任度 社会相似度 社交网络 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]