A ranking method for social-annotation-based service discovery

Abstract

With the rapid growth of Web services, service discovery becomes an important and difficult issue. Traditional UDDI-based and WSDL-based methods of service discovery have low precision, and semantic-based service discovery methods are usually inefficient and time-consuming. We observe that social annotations can optimize both precision and efficiency of service discovery. In this paper, we propose a social-annotation-based service discovery method by using a learning to rank method, and propose two algorithms, Query Annotation Relevance (QAR) and Service Annotation Ranking (SAR), to calculate the dynamic Query-dependent feature and the static Query-independent feature respectively. Our experiments show that our method is effective for improving service discovery performance.