COMPUTATIONAL TOPIC TRUST WITH USER’S INTERESTS BASED ON PROPAGATION AND SIMILARITY MEASURE IN SOCIAL NETWORKS

Abstract

The purpose of this work is to present a model of trust computation being defined by means of the context of user’s interests and degrees of direct interaction among peers. A dynamic structure of social network based on a hierarchy of peers into k-layer neighbors w.r.t. a given source peer is investigated. We introduce a similarity measure being constructed from the space of vectors of users’ interests. Up on the location of a target peer w.r.t. the hierarchy, we propose methods of estimating trust values via propagation as well as the similarity. Some classes of functions for supporting the computation of trust values in these cases are also
considered.