Research On Personalized Web Service Recommendation Based On Sparse QoS And Collaborative Filtering

The rapid development of web makes the information explosion.Recently,more and more algorithms occur to arise the improvement of web service recommendation system.The key problem of web service recommendation is the completion and prediction of sparse quality of service.The web service candidates which provide similar functions are hundreds,so the first problem of service recommendation is completing the sparse quality of service,which makes the strange services accessible.The second problem of service recommendation is identifying the personal requests of different users.The accuracy of service recommendation relies on the accuracy of personalized request discovery.Exploring the personalized requests of users to make accurate QoS prediction becomes the most important thing of web service recommendation.Firstly,with the purpose of explore the personalized request of users,this paper proposes User's Preference based Collaborative Filtering(UPCF).The algorithm extracts user's preference from the QoS data and makes the preference data as the main data in the similar neighbors selection process and neighbors priority identification process of users and services,then makes the QoS prediction for web service recommendation.Secondly,the paper proposes Collaborative User's Preference based Collaborative Filtering(CUPCF).The algorithm extracts user's preference from the QoS data and collaborates the preference data and QoS data in the similar neighbors selection process and neighbors priority identification process of users and services,then makes the QoS prediction for web service recommendation.Finally,Location Sensitive and Collaborative User's Preference based Collaborative Filtering(LSCUPCF)is proposed.This algorithm collaborates the personalized location feature and preference feature of users in the similar neighbors selection process and neighbors priority identification process of users and services,then makes the QoS prediction for web service recommendation.The experimental verification is carried out in the WSDREAM database.The final experiment results show that the algorithms proposed can make effective QoS prediction for web service recommendation.