Detecting Malicious Feedback Rating using Similarity Computation for Measuring Web Service Reputation

In the field of service computing, reputation of a Web service is usually calculated using feedback ratings provided by service users. However, the existing malicious ratings and the changing preferences of different service users often lead to a bias towards positive or negative ratings. A reputation computation approach is proposed for accurately measuring the reputation of Web services. We first detect malicious feedback ratings using similarity computation and then adjust the malicious rating with the average rating for computing the actual service reputation. Extensive experiments have been conducted based on a real world Web service dataset to indicate that the proposed approach significantly improves the quality of the recommendation results compared with existing methods. It can also enhance the reliability of service selection.