Research Of Dynamic Trust Model Based On Bayesian Network

Along with the convenience brought by Internet, shopping online is gradually deep into people's lives, E-Business gets rapid development. At the same time, the development of e-commerce is seriously affected by various security risks and ethical factor. Consumers have got kinds of choices and are usually overloaded by large amount of information; as a result, it is hard for them to find the honest trader and qualified goods. The credit and security issues become bottlenecks that restrict the development of E-Business. t is no doubt that, the"trust thing"plays very important role in E-Business system.In this paper, author proposed a dynamic trust management model based on the Bayesian networks, which is built on the basis of the trust management models existed and the characters of Bayesian networks. This model is designed against the problems exist in the trust valuation model used in the C2C E-Business. In the process of constructing this model, we first record the value of each character of the Bayesian networks, then update the trust evaluation, and finally adjust the evaluation with weighted average method. The advantages of this model are as follows: the valuation system is fair and objective the indexes of valuation are reasonable and comprehensive; the combination of Bayesian networks and weighted average method makes use of advantages from both sides, provides preciser and more objective guidance, and is more practical for consumers; the unfair factors are eliminated by recommended trust value; the malicious evaluations are kind of inhibited in detail.Articles include:(1) Author study the trust and the Bayesian network, and make a new dynamic trust model on Bayesian network.(2) Based on the theory of Bayesian networks, author accords to the dynamic trust model has been established in the C2C E-Business environment, and make a new dynamic trust model BNTM on Bayesian network, then analyze the key parameters. Some important but have yet satisfied specific factors have been make and analyze in detail. Solutions have been put forward to. Survey results and a priori knowledge or subjective probability are combined, And proved it's effective and reasonable in theory. The model achieves the design goals Expected.(3) Simulating in MATLAB6.0, the new dynamic trust model has been validated. It is objective and fair, and has comprehensive and effective parameters. Especially eliminates unfair factors for new users, further proof the model rationality and effectiveness.