Trust Evaluation Model Based on Statistical Tests in Social Network

Abstract

A recommendation model is important in the trust environment when the trust between some nodes was lacked or incomplete. Thus the trust evaluation before and after any interaction or recommendation becomes a very important issue to overcome distrust and fake recommendation challenges and help in making decisions. The recommendations are one of the most widespread tools to improve trust, where they can be used for developing a trust model when the performance of the trust model depended on the quality and type of the relations. This paper presents a trust evaluation model based on some statistic tests, which aims to compute the ratio between recommendation to trust, and hence filter out noise recommendation and obtain more accurate and trust values.