Mining and clustering service goals for RESTful service discovery

摘要: In recent years, RESTful services that are mainly described using short texts becoming increasingly popular. The keyword-based discovery technology adopted by existing service registries usually suffers from low recall and is insufficient to retrieve accurate according users’ functional goals. Moreover, it often difficult for users specify queries can precisely represent their requirements due the lack of knowledge on desired functionalities. Toward these issues, we propose a approach leveraging goal (i.e., functionality) mined services’ textual descriptions. first groups available into clusters probabilistic topic models. Then, goals extracted descriptions also clustered based modeling results services. Based clusters, design mechanism recommend semantically relevant help refine initial queries. Relevant retrieved matching user selected with those candidate To improve goal-based approach, further hybrid integrating two approaches. A series experiments conducted real-world crawled publicly accessible registry, ProgrammableWeb, demonstrate effectiveness proposed