A Clustering Based System for Diversifying WSRec Results

Abstract: The use of Web services for various applications has led to the growth of web services on a large scale. Due to the increase in usage of web services, it has become of prime importance to design systems for effective web service recommendation. In our project, we propose a system for effective web service recommendations incorporating users� preferences regarding quality and diversities amongst web services. Users� requirements are considered and mined from his usage history. Then we find functional similarities using clustering techniques followed by applying a ranking algorithm to list top-k services. To discover high quality Web services, a number of QoS models for Web services and QoS-driven service selection approaches have been proposed in the service computing field. In this system user explicitly specifies his/her interests and QoS requirements, and submits them to the service discovery system. Then the service discovery system matches the user�s interests and QoS requirements with corresponding attributes of Web services, and returns those with the best matching degrees to the user.

Keywords: Web Service Recommendation, ANN, feature evaluation, ranking, top-services.