The Research On Ontology-Based Web Services Matching

Web service is the latest development of web technology. It can integrate many services on the net with seamless connection and communication. It can also establish a loosely coupled web applications on the web. It becomes a challenging task to correctly and efficiently find the necessary services from the large amount of Web Services.Traditional service matching technology is achieved via keyword matching. Due to the lack of semantic description, it can't use an intellectualized way to achieve business relations, agreements and the rules of business, resulting in precision and recall rate unable to meet the user's requirement. Most currently researches on concept similarity computation are based on "subClassof" relationship between concepts. However, it does not utilize the concept of semantic information completely. This method also has problems with the imprecise deduction and weak discernment.In allusion to the limitation of traditional service matching algorithms, this thesis extends and improves it. The new algorithm is based on the semantic of web service's input and output parameters. It calculates the similarity of input and output parameters in different ways. The computation of output parameters focuses on improving the precision rate of searching, and the dealing with input parameters focuses on improving search recall rate. Different subClassof relations are given different weights when the semantic distance is calculated. While calculating the semantic similarity between concepts, other relationships are treated other than subClassof relationships. Finally, the similarity of input and output are given different weights, and then we will get the final results of the Semantic Web services matching.This thesis implements the proposed algorithm above. The experimental results show that this algorithm has increased the rate of recall and precision. The distinction of results has also been improved.