PATH ALGEBRA FOR TOPIC TRUST COMPUTATION BASED ON REFERENCES OF USERS IN SOCIAL NETWORK

Abstract

Topic trust in social networks is defined based on combination of interaction experience and user interests. While interests are introduced to classifying users into topics, interaction among users is modeled as a directed graph with weights being a number of connections between nodes. However, such a computation is mainly based on direct interaction being available among users. When direct connections among partners is unavailable, the above model of computation is not applicable. In this paper, we first present a reference topic trust model based on path algebra for topic trust in the context of lacking connection between users. Then we describe an algorithm for estimating such topic trust values.