Research On Semantic Web Service Discovery Method Based On Interface Matching

With the popularization of Web services applications, how to search the user required services quickly and accurately in the large-scale service registered library has become a research hotspots, for which many researchers introduced the Semantic Web to improve service search effect, however, the existing semantic Web service discovery algorithm is still much to be desired. On the one hand, many researches only considered the semantic similarity of interface name, from which separated and ignored the data structure similarities of the input and output parameters. Users'requests could not be satisfied simultaneously in the two similarities. On the other hand, the high time complexity of the semantic algorithms leads to the inefficient of the service, which wastes much time by matching the unconnected service with every request and database. This research mainly explored the above two questions.Firstly, one more accurate algorithm of similarity of distance semantic-based was conducted in this study, which improved the existing algorithm of concept similarity by calculating the semantic distance, depth and density of the concept nodes in the ontology tree. Moreover, one multi-keyword similarity calculation also was adopted to describe the similarity of the service text.Secondly, the comprehensive similarity calculation method was implemented to determine the similarities of interface concept semantic, parameters data type and data constraint condition, which be valued by their different effects to the interface similarity.Finally, a method of semantic web service discovery based on the interface matching was applied to save the users'time. The inverted index structure in the service registry library was established according to the output interface, presented the information of meeting the users'demands and avoiding the unrelated mismatching. Following this useful information, the method got the service list which ordered the service similarity, then returned to the users'requests threshold.The result indicated that the semantic web discovery method based on the interface matching not only increased the precision ratio and recall level, but also make the service discovery faster. Data analysis showed that the greater the service registry library was the higher efficiency the method had.