MODELING COMPUTATIONAL TRUST BASED ON INTERACTION EXPERIENCE AND REPUTATION WITH USER INTERESTS IN SOCIAL NETWORK

Computational trust among peers plays a crucial role in sharing information, decision making, searching or attracting recommendations in intelligent systems and social networks. However, most trust models focus on considering interaction forms rather than analyzing contexts such as comments, posts being dispatched by users on social media. The purpose of this paper is to present a novel model of computational trust among a truster and a trustee in two stages. First, we construct a function, named experience topic-aware trust, whose computation is based on users interaction and their interests on topics. Then we establish a composition function, named topic-aware trust, which is constructed from the estimation of truster’s direct experience trust and some reputation trust on some trustee. Our experimental results show that the interest degrees affect on trust estimation more than interaction ones. In addition, the more interest degree in a topic users obtain, the more trustworthy they are.