Research On Fusion Trust Model Based On Virtual Community

With the popularization of the internet and the development of technology, the traditional physical communities become unable to meet the increased demand for information exchange. Virtual community is a free and online community opened for everyone. Because it is convenient to exchange information, the number of its users has increased in these years, and it gradually become a new trend. But virtual communities are also has some risks. The anonymity, openness, and the low trust virtual communities increased the difficulty of community management. In transaction-based virtual community, the existing trust management mechanism can provide security for the good transaction, but in other types of virtual communities, the trust mechanism run too subjective. So, now we need to analyze and improve trust management in non-transactional virtual community.This dissertation presents the current problems in the virtual community, analyzing the concept and characteristics of trust and credibility, and trust issue in virtual community environment.In order to solve the risk problem in non-commercial virtual community, which is caused by the difficulty to obtain the trust relationship between users and the different trust in vary domain. This dissertation introduces the concept of the domain. Take this concept which is come from the requirements engineering analysis methods fused into the community trust model. It presents a method based on trust implicit feedback to obtain direct trust; the overall credibility, domain credibility, clipping algorithm is proposed to get indirect trust, these algorithms can ensure the accuracy and reliability of the indirect trust.In order to solve the incompleteness problems of domain division and problems in the domain of deception, a classification method based on user domain VDKNN virtual community to identify the user's home domain is proposed. The malicious behavior detection and prevention based on areas of maturity and capacity factor of the domain factor are proposed.In order to maintain the objectivity of the virtual community trust, we should to avoid subjective factors undermine trust mechanism and interference by malicious users. So we research the mechanism of risk and build trust fusion model.The Experiment simulate the credibility changes of two different types of users in the community and take the fused model compare with EigenTrust model. The validity and rationality of the proposed trust model are verified by the result of this experiments. Finally, the dissertation further verify the effectiveness of the model by simulation community scenes.