Estimating the reputation of newcomer web services using a regression-Based method

In this paper, we propose a novel method to estimate the initial reputation values of newcomer web services. In fact, the reputation of web services is one of the criteria used for recommending services in service-oriented computing environments. The lack of evaluating the initial reputation values can subvert the performance of a service recommendation system making it vulnerable to different threats like whitewashing and Sybil attacks, which negatively affect its quality of recommendation. The proposed method uses Quality of Service (QoS) attributes from a side, and reputation values of similar services from the second side, to estimate the reputation values of newcomer services. Basically, it employs regression models, including Support Vector Regression, in the estimation process of the unknown reputation values of newcomers from their known QoS values. We demonstrate the efficiency of the method in estimating the reputation of newcomer services through statistical evidences gathered from experimentations conducted on a set of real-world web services.