Web Service Recommendation Methods Based On Collaborative Filtering

As the requirements for Web service application become more complex, single Web service can’t meet users’ requirements. As a result, composite Web service, which is value-added service of Web service, composites single Web services according to some business logic so that it can meet users’requirements at a higher level. Owing to that, redundancy has become prevalent with many service providers offering same or similar services. So, how to select or recommend the most suitable Web services for users is becoming a hot issue in research field.To address the service selection problem of each task, the main issue of this thesis is: How to recommend the most suited services for tasks from the services that meet users’SLA requirements.This thesis proposes Web service recommendation methods based on collaborative filtering,which mainly use the historical execution information of composite bussiness instance to recommend service. It considers the personalized information of the user and uses the information to finds.the similiar historical instance information which then, is calculated to recommend the suitable service for the user.This thesis first proposes a Web service execution model. The new model, which is different with previous theoretical models, contains not only QoS characters but also environmental characters, input characters and time characters. So the model can take a account of other users’experiences more accurately. Then, in order to better predict the QoS of Web service which users experience, this thesis puts forwards a QoS collaborative filtering based prediction method, including using other users’experience on the service and the historical time of the experience as a factor in calculating user similarity. After that, this thesis focuses on using the BP neural network to address the sparse data problem which collaborative filtering technology generally faces. In the end, the thesis introduces the application of the recommendation mechanism in detail and represents the using experiments of this mechanism including the analysis of the accuracy of this mechanism.