Research On Reputation-Based Trust Modeling

Nowadays the rapid development of e-commerce, with the network platform, isbecoming an important emerging business model. But at the same time, In such anopen, dynamic, anonymity environment, e-commerce transactions are flexible, but fullof risk, which makes e-commerce trust problem is more serious and becomeconditionality factors in the development of electronic commerce. So how to establishmaintain and promote e-commerce transaction between the two sides of the trustrelationship, reduce the e-commerce transaction risk, reduce transaction costs,improve the safety of transactions in the development of e-commerce has became animportant problem to be solved. While building trust model can effectively solve theproblem of trust and help both parties to establish a reliable trust relationship tosecurity.This paper is based on trust mechanism research and aiming to the application.At first, we sort and classify the empirical findings on the e-commerce trust modeland with the comparison of the e-commerce trust evaluation model, we supplementthe trust evaluation model factors. We segment the trust model into three type basedon research methods, they are weighting models, probabilistic models and fuzzy logicmodels. With considering the factors of trust selection, quantification and modelingmethod based on trust, we make some improvements on the basic of Jiang’s trustmodel. We improve the local trust model and global trust model, introduce therewards and punishment mechanism and enrich the factors of trust model. At last, wedesign a simulate experiments to verify the accuracy, anti-attack capability of theproposed model. By simulation experiments, we find our model have a higheraccuracy and anti-attack capability.