A General Self-Adaptive Reputation System Based on the Kalman Feedback

Abstract

With the rapid development of the web services, e-commerce and social network applications, a robust reputation system to establish trustworthiness between mutually unknown entities is becoming increasingly important. This paper proposes a general self-adaptive reputation model, which uses the weight factor of each feedback to inherently support the defense of fake feedbacks. Moreover, we design a reputation system by using the improved Kalman Filter based on the factor of weight. With this method, we can not only get an accurate prediction for the service provider, but also resist malicious feedback attacks. Our reputation system is proved to be more robust and accurate compared with the traditional methods in the simulation and experiment.