Abstract:

Based on subjective Bayes method, we predicted user trust relation in social networks. The method utilized the existing trust relationships and rating relationships between users, took a rating record from user to another user as evidence and calculated the quantity representation of “trust” and “distrust” by merging the evidences based on a rule set obtained by supervised learning, thus predicting and spreading the trust relationship among users by using multi-valued logic method. The comparative experimental result shows that the proposed method can solve the usertrustprediction problem in social networks with high accuracy.

Key words: trust relationship, subjective Bayes method, social network, uncertainty reasoning