Research On Association Analysis-Based Web Service Matchmaking Strategies

Web services with loosely coupled, language-independent, platform independent features have become a major method of opening and provisioning information, services and business ability in Internet computing environment. It is an important task for business developers and users to find suitable Web services to meet their business development needs quickly and accurately in so many resources. Semantics is the key to enhance the level of automation and intelligence of Web services applications. The semantic of Web services reflects service functions, and semantic association reflects the relationship between the service functions, which have close contact and interaction. Association analysis is a form of searching the intrinsic link from the large amounts of data. If semantic association analysis can combine with the Web service matchmaking, it will improve the accuracy of the matching effectively.This thesis focuses on Web services matchmaking strategy based on the association analysis and carries out in-depth research, which makes the following research results.(1) By using semantic information in the service description, the concept of semantic association between services is proposed, which is adopted to extend the existing matching algorithms by using the semantics matching method based on the calculation of association matchmaking degree. The main principle of the method is that, in order not to give up a potential match, the combining services which are eliminated in the one-to-one matching to achieve a one-to-many service matching through semantic association. Further, the matchmaking degree of the Web service in the obtained candidate services and the candidate services set are both calculated, and then the service of the best matchmaking degree by sorting is obtained. The implementation framework of the method is given to illustrates the process and characteristics of the method with an example.The proposed method is compared to the existing methods by using simulation in terms of time cost, the recall rate and precision. The experimental results show that compared to the existing semantic Web service matching methods with single interface parameters, the proposed method can provide more accurate service matching results in the premise of increasing reasonable time overhead.(2) A fine-grained Web services matching algorithm based on semantic association graph is proposed. On one hand, analyses services’OWL-S process model and services’multiple executions to cope with the complex service matching case and to implementing the fine-grained matching. Using semantic association graph can capture the relationships among the matched services better, and the proposed algorithm takes into account composite process of any structure. On the other hand, the extended OWL-S service profile is used to filter matching, in order to reduce the number of candidate services in the constructing graph stage, thereby the service matching efficiency is improved. Finally, an example is given to illustrate the effectiveness of the algorithm.