Research On Trust Model Of P2P E-commerce

E-commerce has become research hotspot, but the problems of bandwidth and the increasing requests for server block the development of e-commerce. Recently, P2P e-commerce is developing step by step as a new model of e-commerce. It is important that the problem of trust must be solved whether in e-commerce field or in P2P network. Trust model provide a way to solve this problem. The main functions of trust model are evaluating the trust relationship between entities, computing the trust value and so on. So, we put forward two kinds of trust model used for P2P e-commerce with different math methods. Both of them can evaluate trust effectively, compute the trust value accurately and stop the vicious activities. In the research of P2P e-commerce trust model, we have done the work below:(1) Analyze the trust models that have existed in P2P, and summarize their advantages on trust expression, evaluation and computing, also their disadvantages. According to the math method they use, we classify them into two kinds: trust model based on certain theory and trust model based on subject uncertain theory.(2) Put forward a trust model based on polling logic. Trust is composed of direct trust and indirect trust. In the computing of direct trust, we bring in five factors including transactions context, feedback in terms of amount of satisfaction, number of transaction, time of transactions and nicety of polling. After that it can stop the malicious activities including cheating, imputation and so on. it can also solve the problem of FreeRider which exists in P2P, and stimulates peers to poll accurately. In this trust model we provide a way to collect recommends from other peers, which uses the voting logic. It not only ensures the security, also reduces the complexity of reputation computing. After the experiments, we validate the capability of preventing kinds of attacks, and prove the advantage compared with other models.(3) Put forward a trust model based on fuzzy set. It's a method based on subject uncertain theory. Adopting the fuzzy comprehensive evaluation, we compute direct trust value more diversiform and flexible. We bring in the punish factor, encouragement factor and time factor in the computing which can punish the malicious peers and encourage the good peers.(4) Put forward a method to update the recommend trust. In the formula we bring in the similarity coefficient. Using this method, we can not only find the peers who...