A calculation method for social network user credibility

Fang Zhou, Jian-Lin Jin, Xiaojiang Du, Bowen Zhang, Xu-Cheng Yin

Abstract: Trust plays an important role in helping social network user make appropriate decision from abundant social information about products. A variety of social review information (e.g. ratings, voting and tags) emerge with the huge number of products on the Web. How they are utilized for searching and finding appropriate item is investigated. In this paper we propose a user credibility calculation method, which calculates user's credibility through his previous review information. By evaluating facets of every review, an integrated numerical value which denotes the reviewer's credibility can be calculated. This value can be used to rank products, further to help user making appropriate decision. Experiments on social book search database show that this calculation method improves effectively accuracy of recommended items.