Research On The Predictionof Membership Degree And Distribution Of Weights In Fuzzy Trust Model Of P2p

At present,peer safety issue in P2P network has become a hot research.There have been a number of trust models.The evaluation of trust is the most importment one in trust model.In the dissertation, trust evaluation method in existing trust model is studied.The way of simple calculation of membership and random distribution of the weight factors is improved in most trust evaluation method.On the base of fuzzy of thinking, the formula calculating the weight factor of direct trust and recommended trust is defined; the forecast model of membership degree of factors impacting entity trust by time-series algorithm is established in P2P network and predictive value of membership of trust factor is obtained; The data of acquisition of network bandwidth, resource value, storage capacity, size of upload file and size of down load file is Analysized and weights of factos of entities trust are acquired by Fuzzy Clustering and Rough Set Algorithm. The simulation results show that prediction of membership degree based on time series algorithm is more effective than direct calculation of membership degree and the dynamic changes of membership is also reflected. Distribution of weight factors based on Fuzzy Clustering and Rough Set algorithm is more accurate than random allocation method in the evaluation of entities trust.