Internet-Based Evaluation and Prediction of Web Services Trustworthiness

As most Web services are delivered by third parties over unreliable Internet and are late bound at run-time, it is reasonable and useful to evaluate and predict the trustworthiness of Web services. In this paper, we propose a novel approach to evaluate and predict Web services trustworthiness using comprehensive trustworthy evidences collected from the Internet. First, we use an effective way to collect comprehensive trustworthy evidences from the Internet, which include both objective evidences (e.g. QoS) and subjective evidences (e.g. reputation). Second, Web services trustworthiness is evaluated with collected trustworthy evidences on a regular basis. Finally, the cumulative evaluation records are modeled as time series, and we propose a multi-step Web services trustworthiness prediction process, which can automatically and iteratively identify and optimize the model to fit the trustworthiness series data. Experiments conducted on a large-scale real-world dataset show that our method manages to collect comprehensive trustworthy evidences from the Internet and can effectively evaluate and predict the trustworthiness of Web services, which helps users to reuse Web services.