Trust propagation and trust network evaluation in social networks based on uncertainty theory

Abstract

Uncertainty distributions can help resolve the difficult problem of measuring subjective uncertainty within a trust relationship. This paper studies trust propagation and trust network evaluation in social networks by using uncertainty theory. First, we identify types of relationships between decision-makers (DMs) and construct the underlying trust network by defining the correlation function based on uncertain distances. Second, uncertainty optimization models of single-path and comprehensive indirect trust are developed so that the comprehensive indirect trust value between DMs can be simply calculated. A maximum belief degree model is introduced to compute the maximum belief degree and to obtain the optimal trust propagation path between two DMs. Third, by defining such a concept as consilience degree of a trust network, the trust relationship between DMs can be effectively measured. We also evaluate a trust network respectively from the perspectives of individual influence, the consilience level of the decision group and the stability of the local trust network. Finally, a real-world case of selecting the members of an enterprise credit group is illustrated to confirm the validity of our proposed methods and concepts in this paper.