Research Semantic Web Service Matching Based On Similarity

With the rapidly development of internet and application, Web Services as an emerging model of web applications and remote access standard, has a very good solution to the problem of data and application's integration and sharing on heterogeneous platforms. With the increase of web Services, finding the services accurately and efficiently from the mass of web services to meet the demand has become increasingly difficult. Traditional web services discovery mechanism based on keyword matching lacks of machine-understandable semantic information, which leads to low recall and precision and can't achieve intelligent search. Therefore, the introduction of machine understandable semantic information in semantic-based basis provides more accurate and efficient match making algorithm of web services and mechanisms to quickly find relevant web services has become an issue.This essay first briefly introduces the related concepts of ontology and OWL focuses on OWL-S, which is the bridge that connects web service and semantic ontology. A lot of researches about the semantic web service matching are based on OWL-S. The content of this essay is also OWL-S based.Then, The deficiency of services distinguishing capability in web services matchmaking approaches, which is based on description logic reasoning, is analyzed. A simple and efficient calculation method for semantic distance is presented. Based on this calculation method, a new web services matchmaking approach with enhanced semantic precision is proposed under the consideration of matchmaking efficiency and precision. On the basis of description logic reasoning, the Proposed matchmaking method introduces semantic distance calculation to enhance the services distinguishing capability of matchmaking system.At last, this thesis designs a multiple level service matching framework. The framework takes advantage of similarity algorithm which is proposed by this paper to implement matchmaker, and make use of service ontology to test matching effect of framework.