On Trustworthy Reputation Evidence Establishment: Rating Analysis and Defense against Dishonesty

The guarantee of high trustworthiness holds the balance for secure sharing and efficient collaboration among entities in wide distributed, dynamic domain. Recently, most papers elaborate on architectures or mechanisms for designing reputation service and resource selection. The presence of inaccurate testimonies and malicious referrers is necessary to be considered. Based on D-S evidence theory, guided by rating reliability in psychology, the proposed approach can filter out dishonest feedbacks and respects entity's rating habits at the same time. Experiments show the cost of calculation is low. It is apt to deploy in open environment and will be effective after short time of training.