Research On Context-aware Web Services Recommendation

With the development of Web service technology, relevant products and technologies of Web Services are getting more and more mature. Meanwhile, more and more available Web services on the internet are provided and similar or even the same functional services also appear. With so many services, it becomes more and more difficult for users to choose a service meeting their needs. Aiming at solving these problems, optimization recommendation algorithms are used in Web services selection process. In these algorithms, QoS information is considered as important criterion in selecting Web services. To some extent, these algorithms improve the quality of service selection. But they are difficult to satisfy user’s personalized demand. To solve this problem, context-awareness technology is introduced into service recommendation. In service recommendation process, through analyzing contexts of user behavior, user preference and QoS information, the best service list suitable for current user will be selected.This paper considers context information of user’s history behavior, user’s current behavior and QoS properties and presents four service recommendation algorithms suitable for different context situation which are user based collaborative filtering recommendation algorithm, service based collaborative filtering recommendation algorithm, QoS based service recommendation algorithm and time-weighted based service recommendation algorithm. Meanwhile, recommendation algorithm selection strategy based on different context information is also provided. In service recommendation process, context information has two roles. The first is that they determine the selection of service recommendation algorithm. The second is, as the important parameters of recommendation algorithm, they directly influence the service recommendation result. This context-aware Web service recommendation algorithm selection and execution strategy not only realizes the personalization of service recommendation algorithm selection, but also realizes the personalization of service recommendation results. Finally, the service recommendation algorithms presented in this paper are implemented and context-aware web services recommendation system is also designed and implemented. Experiment analysis shows that the service recommendation algorithm presented in this paper can improve users’ satisfaction to service selection from three aspects which are precision rate, recall rate and coverage rate of recommendation results. The research in this paper has some theoretical and practical application value.