Trust-Based Personalized Service Recommendation: A Network Perspective

Recent years have witnessed a growing trend of Web services on the Internet. There is a great need of effective service recommendation mechanisms. Existing methods mainly focus on the properties of individual Web services (e.g., functional and non-functional properties) but largely ignore users' views on services, thus failing to provide personalized service recommendations. In this paper, we study the trust relationships between users and Web services using network modeling and analysis techniques. Based on the findings and the service network model we build, we then propose a collaborative filtering algorithm called Trust-Based Service Recommendation (TSR) to provide personalized service recommendations. This systematic approach for service network modeling and analysis can also be used for other service recommendation studies.