Context-Based Web Service Reputation Measurement

Reputation of Web service plays an important role in selecting the best service from massive services. Traditional approaches of measuring Web service reputation are based on all the historical ratings. Although these approaches are effective in the same context environment, they fail in obtaining accurate reputation measurement when users´ context is different. In this paper, we propose a context-based Web service reputation measurement approach by weakening the effects on each user´s rating caused by different context. The approach first classify ratings based on users´ context. Then the approach calculates the difference in ratings given by one user with different context to get users´ inner rating. Finally, we adopt User-based Collaborative Filtering to measure the reputation of each Web service. We have implemented our approach with extensive experiments and the results show that our approach outperforms other approaches.