Research On The QoS Service Selection Based On The Historical Choice Information

With the rapid development of Web services technology, Web services are becoming more widespread and the number of Web services on the Internet is growing. So, the research of how to quickly find and choose the Credible Web services that meet the functional needs of service consumers and its QoS needs in numerous Web services have become an increasing concern.In the research of Web service discovery and selection, the recall and precision rates of the traditional SOA model based on keyword matching are relatively low. Therefore, Semantic Web technology in combination with Web services by the researchers. However, at present, most of the semantic-based QoS service selection method just consider the matching of service QoS values, not the QoS semantics. In addition, in the quantification of service quality values, the general approach is to take the average value and the information is "loss". And on the other hand, a lot of historical choice information is no reasonable application.In the credibility evaluation of Web services, most of the existing methods is according to the direct credibility based on historical records and the recommended credibility based on acquaintance recommendation. Those methods don't take into account the context when service consumer make the service evaluation, the user's own credibility and the impact of user's preferences on the credibility.This paper studies the issues mentioned above and presents a QoS service selection model based on the historical choice information(WSDM-HQ). This model realize the QoS service selection algorithm based on historical choice information and the Web service credibility evaluation algorithm based on virtual community. In the process of Web service discovery and selection, the property values of QoS, in this paper, quantified into intervals. And then, according to intervals' similarity of old and new service requests, we find the service which meet customer's needs from the existing service selection information. And if we can't find anything, we find the service which meet customer's needs from the UDDI Registry. This method increases the speed of Web services selection. Besides, in the process of service matching, this method consider the basic description of the service, IOPE, QoS semantics and QoS values four aspects to improve the accuracy of service discovery. After the service call completed, we evaluate the credibility of services called in this time. The evaluation algorithm is Web service credibility evaluation algorithm based on virtual community. The first, the service consumers with similar expectations is composed consumers virtual communities. And then, the service providers with a certain degree of similarity is composed service providers virtual communities. In that way, we can calculate the credibility of Web services according to the historical experience of consumers in consumers virtual communities and the similarity between consumers and service providers. After that, we can generate corresponding historical records or revise the effectiveness of existing historical records. Finally, we carry out some simulation experiment, the results of simulation experiment indicate that this model is feasible and effective.