PROJECT TITLE:

Diversifying Web Service Recommendation Results via Exploring Service Usage History - 2015

ABSTRACT:

The last decade has witnessed a tremendous growth of Net services as a significant technology for sharing data, computing resources, and programs on the Net. With the increasing adoption and presence of Internet services, style of novel approaches for effective Internet service recommendation to satisfy users’ potential necessities has become of paramount importance. Existing Internet service recommendation approaches mainly specialise in predicting missing QoS values of Internet service candidates which are interesting to a user using Collaborative Filtering approach, content-primarily based approach, or their hybrid. These recommendation approaches assume that suggested Web services are independent to every different, which sometimes could not be true. Hence, many similar or redundant Net services might exist in a recommendation list. In this paper, we propose a novel Internet service recommendation approach incorporating a user’s potential QoS preferences and diversity feature of user interests on Net services. User’s interests and QoS preferences on Internet services are first mined by exploring the Net service usage history. Then we tend to compute immeasurable Internet service candidates by measuring their relevance with historical and potential user interests, and their QoS utility. We tend to also construct a Web service graph based mostly on the functional similarity between Net services. Finally, we present an innovative diversity-aware Internet service ranking algorithm to rank the Net service candidates primarily based on their scores, and variety degrees derived from the Net service graph. Extensive experiments are conducted based mostly on a true world Web service dataset, indicating that our proposed Net service recommendation approach considerably improves the quality of the advice results compared with existing strategies