Abstract: In the current e-commerce industry, users need to check for the historical evaluation of the commodity before buying goods mostly, but false recommendations and massive redundancy recommendation problem which widely exists at present seriously affect the effectiveness and accuracy of users’ obtaining recommendations. For getting trust in e-commerce transactions, an e-commerce trust model SNTrust based on social networks is proposed. In SNTrust model, each user maintains a trusted social network set; a topological structure of the user information diffusion in the social network is described in the model, and the recommendation node trust is introduced to measure its credibility, the algorithm to compute the trust value of a transaction is given. Simulation results show that the SNTrust trust model can effectively improve success rates of e-commerce transactions, and curb the malicious behavior to a certain extent.

Key words: e-commerce, social networks, trust computing, recommendation