A Web service selecting method based on the prediction with QoS Parameters

In order to enable Web service consumer acquiring the QoS value as a reference for service ranking and sorting, a Web service ranking algorithm based on the QoS parameters prediction is proposed. First, according to the measure of Service Quality Conformance, the service quality reports were produced, the creditable reports were marked, and the preprocessing method was used to exclude those cheating reports. Second, the QoS value was predicted using the fuzzy regression method, whose I/O parameters were the property and evaluation value for service quality respectively, so the fuzzy coefficients can be determined and the QoS evaluation was hazed with Symmetric triangular fuzzy number. Finally, the results were employed as the service selecting basis. The experiment results demonstrated that the predicted value was close to the real QoS value gradually as the sample number of the service quality report was increasing. The service quality prediction method compared to the others in the service discovery enabled the average R-Precision error to be controlled within 2.7% with varying the parameter of Service Quality Conformance and the QoS difference in service advertisement and requirement accepted by service consumer.