The Recommendation Mechanism Based on Trust Network for E-Commerce

Abstract

In Internet-based e-commerce, the transaction entities’ trust levels provided by credible recommendation can be viewed as a key indicator for users selecting transaction partners. The trust of transaction entities objectively and comprehensively represents the integrated influence of various trust attributes. Aiming at vicious recommendation in e-commerce, the concept of adopt coefficient is introduced, and the trust vector of the transaction entity is defined. The average amount of each transaction, the transaction number, the bad ratio of target entity is identified as the attributes of the trust feature vector. The trust network for e-commerce is established, and corresponding trust calculation algorithm is proposed and described in detail. According to the model, a user selecting transaction partners can make smart decision considering the potential partners’ recommendation trust and direct trust. The experimental results illustrate the effectiveness and availability of the trust network model for e-commerce and trust calculation algorithm.