AUTrust: A Practical Trust Measurement for Adjacent Users in Social Networks

Abstract

Trust is the foundation of interpersonal relationship and interaction as in social networks as in real world. Previous researches about the trust between users mainly focus on nonadjacent users in social networks, where the trust between adjacent users is assumed to be known. However, in most social networks, the trust between adjacent users is unknown, that limits the applications of those nonadjacent users trust models. In this paper, we propose an adjacent users trust measurement model(AUTrust) and discuss how to construct a trust social network with AUTrust. The factors affecting trust between users could be classified into three dimensions: (i) similarity between users, (ii) familiarity between users, (iii) users' social reputation. The last two dimensions can be used to reflect asymmetry of trust between users. Then we discuss the methods of quantifying trust with these dimensions. Users' information that AUTrust needs can be collected with API provided by social networks, so the model can be widely applied to social networks. Finally, we evaluated the model on 10-million-level Ten cent Weibo dataset, and the experiment proves the practicability and universality of AUTrust model.