The trust levels of cloud services should be evaluated to ensure their reliability. The effectiveness of these evaluations has major effects on user satisfaction, which is increasingly important. However, it is difficult to provide objective evaluations in open and dynamic environments because of the possibilities of malicious evaluations, individual preferences, and intentional praise. In this study, we propose a novel unfair rating filtering method for a reputation revision system. This method uses prior knowledge as the basis of similarity when calculating the average rating, which facilitates the recognition and filtering of unfair ratings. In addition, the overall performance is increased by a market mechanism that allows users and service providers to adjust their choice of services and service configuration in a timely manner. The experimental results showed that this method filtered unfair ratings in an effective manner, which greatly improved the precision of the reputation revision system.
展开▼
