Trust prediction with trust antecedent framework regularization

Abstract

In recent years, many discipline theories are developed for understanding trust and solving the data sparse problem. Trust Antecedent framework is an integrative and well-known model in management science, which takes ability, benevolence and integrity as three key factors to explain how trust relations are established between a trustor and a trustee. In this paper, we propose a new trust prediction model based on Trust Antecedent framework (TA) and matrix factorization. We focus on how the factors of TA affect user’s trust in online social networks. TA is incorporated into a matrix factorization with a regularization term to enhance the trust prediction performance. Our experiments conducted on a real-word dataset from Ciao demonstrate that our approaches outperform other state-of-the-art methods in trust prediction.