Recommendation with Implicit Trust Relationship Based on Users’ Similarity

Abstract

Recommendation algorithms based on trust relationship have been shown with a great influence. Many existing algorithms only consider the trust relationship. But in the real world, most users have few trustworthy person can be added to the trust list, they do not know or unacquainted each other. So, the implicit trust relationship is valuable by inferring more trust relationships. This paper proposes a new method to mine the implicit trust relationship based on the users’ similarity. The proposed methods combine the trust-based recommendation algorithms and show the effectiveness in the FilmTrust data set.