Web Service Reputation Prediction Based on Customer Feedback Forecasting Model

In the Service Web, customers´ feedback constitutes a substantial component of Web Service reputation and trustworthiness, which in turn impacts the service uptake by consumers in the future. This paper presents an approach to predict reputation in service-oriented environments. For assessing a Web Service reputation, we define reputation key metrics to aggregate the feedback of different aspects of the ratings. In situations where rating feedback is not available, we propose a Feedback Forecasting Model (FFM), based on Expectation Disconfirmation Theory (EDT), to predict the reputation of a web service in dynamic settings. Then we introduce the concept “Reputation Aspect” and show how to compute it efficiently. Finally we show how to integrate the Feedback Forecasting Model into Aspect-Based Reputation Computation. To demonstrate the feasibility and effectiveness of our approach, we test the proposed model using our Service Selection Simulation Studio (4S). The simulation results included in this paper show the applicability and performance of the proposed Reputation Prediction based on the Customer Feedback Forecasting Model. We also show how our model is efficient, particularly in dynamic environments.