RESTful Web Service Matching Based on WADL

Service matching is essential for web services as increasing of demands of service discovery and composition. Many existing researches on Restful service matching try to mine their features such as term frequency and relationship by traditional semantic ways like clustering, pattern analysis and classification trees. However, we think that for Restful services many of these features could be easily obtained in view of application. Based on WADL (Web Application Description Language), the main idea of our matching model is application oriented semantic matching (AOSM). We will first analyze service application features such as functionality or design purpose of different parts of the service description, then give suitable semantic matching strategy based on these features. We think by introducing this analyzing method, the efficiency and precision of semantic matching will be improved compared to only using common semantic ways. Results from experiments demonstrate that the proposed matching method has achieved a precision rate to 86% and a recall rate to 82% for Restful web services we extracted from several platforms.