A novel evidential trust evaluation algorithm

Trust and reputation are being used by commercial reputation or rating systems for online trading communities over Internet. Evaluation of trust for an unknown agent in web environment is a challenging job. In this paper, we propose a novel evidential trust evaluation algorithm (NETEA) for computing recommendation trust by modifying parallelization algorithm [1] suggested for Bayesian social networks. In order to solve the problem, a definition of trust scaling factor (TSF) has been proposed. This is extension to our previous work [2]; here we apply Dempster-Shafer (Evidential) theory for computing recommendation trust. We identify limitations and propose some extensions for reliable trust evaluation using new decision fusion techniques.