A Latent Trust Discovery Algorithm Based on Optimization of Trust Network

Abstract

Trust networks can be divided into user-to-user and user-to-commodity trust subnetworks. In those networks, it can be distinguished two kinds of classical trust relationships such as functional trust and referral trust. In fact, there is a third invisible trust relationship which we call latent trust. We propose the discovery algorithm to find the latent trust and compute the trust value of user-to-commodity trust which can help customers to directly make the final purchase decisions. Experiment results with the extended Epinions’ dataset illustrate that this algorithm has a better performance than other similar algorithms.