A Bayes-based method for trust valuation in electronic commerce

Abstract

With the development of the Internet, electronic commerce has shown a rapid increase, in the open dynamic Internet how to establish a complete model for trust valuation, objectively representing the credits of trading parties and improving the security of online trade, has become the key issue. To solve the problem of behavior credit in e-trade, this paper presents a trust valuation method based on the behavior pattern showed in the process of e-trade. With this method, a Bayesian model for assessment is presented, composed of nodes of behavior pattern matching levels and nodes of credibility attributes; and the grade of the consumption credit index and that of the sale credit index can be inferred. The effectiveness of this method has been verified by an actual electronic trading platform.