Web Service Matching Method Based On Semantic Similarity

With the wide application of Web services, how to find the necessary serviceaccurately and efficiently has became the current study hot spot. The traditionalWeb service matching methods based on keywords obviously lack enoughsemantic information, and the introduction of ontology concept and the raise ofsemantic Web services description language OWL-S, have greatly improved theservice semantic expression ability and discover ability, thereby the machinecan understand and deal with the semantic information, support automaticallydiscovery, and can improve the recall and precision. So, the study of Webservice semantic matching and found have important oretical and practicalvalue.In view of limitations of the current study to the service function matching,combines the knowledge of dynamic description logic reasoning and similaritycalculation, this paper presents a Web service IOPE function matching methodthat base on semantic similarity. Take precondition and effect matching to thesame category service setts in preliminary screening, do describution andreasoning to service by dynamic descripution logic language, and realizematching algorithm in the support of ontology reasoning machine Racer.The ultimate goal of Web service matching is to find the similarity betweendifferent services, the more similar, the more matching, on the contrary, themore not like, the more not matching. Do input and output matching on the baseof descripution logic reasoning and calculate the semantic similarity of servicefunction parameters I/O, considering the two service setts weather come fromthe same ontology, introduce the concept semantic distance, and Propose onecomputing method of semantic similarity that base on semantic distance.For the same ontology, this paper do semantic distance expand from the threeaspects of information contained by directed edge, the directed edge type, andthe concept density, At the same time, introducing the dissymmetry factor,turnning semantic distance to semantic similarity, this method makes the last concept semantic similarity computation more exact. For the differentontologies, this paper conside the affect of the semantic distance, conceptattributes and the number of causes to semantic similarity, and then set differentweights to calculate the semantic similarity between concepts. Thereinto, thedistance similarity calculation is on the base of the distance-based similaritycalculation method, joining the concept of bridge-connected, so as to calculatethe semantic similarity berween the concepts from two heterogeneousontologies.At the base of study, this thesis designs and realizes a multiple level servicematvching framework based on IOPE comprehensive matching. At last, alsomake use of service ontology to test maching effert of framework, and checkingthe feasibility of this presented method.