Applied Mathematics & Information Sciences

Author Country (or Countries)



With the increasing adoption and presence of Web services, how to recommend Web services to users that satisfy their potential functional and non-functional requirements effectively has become an important and challenging research issue. In this paper, we propose an enhanced Web service recommendation approach, named iAWSR (improved Active Web Service Recommendation), that explores service usage history of users to actively recommend Web services for them. In iAWSR, we propose new methods for computing functional similarity and non-functional similarity of Web service candidates, and a hybrid metric of similarity is developed by combining the two sources of similarity measurement. iAWSR then ranks publicly available Web services based on values of the hybrid metric of similarity, so that a top-k Web service recommendation list is generated for the user. We propose an effective overall evaluation metric to evaluate our improved approach. Large-scale experiments based on real-world Web service datasets are conducted. Experimental results show that iAWSR outperforms the existing approach AWSR on Web service recommendation performance.

Digital Object Identifier (DOI)