SVM-Based Swift Trust Rating Model in E-commerce

It is a long time since researchers have devoted themselves to trust and reputation rating in E-commerce, and some useful fruits have also been achieved. However in my opinion, there should be a long way to go before human really master the trust mechanism. This paper first investigates swift trust rating indexes, which include interdependence, intensity of focusing role, categorization, environmental factor and power of action and so on. And after they select RBF as the K-function, the authors make experiments in LIBSVM environment based on the collected sample data. Lastly they get efficient support vectors. The verification results show good consistence with the facts, and the accuracy of classification is high enough to verify the fitness of using SVM to model swift trust rating.