Research On Active Service Component Library Based On Ontology And Component Ranking

As the development of Internet and the improvement of enterprise e-business, more and more organizations and enterprises are gradually shifting their business models to Internet, which results in many more applications on Internet. By learning and using the open interconnecting standard on Internet, web service builds a generic, platform-independent and language-independent technical specification on the basis of varieties of heterogeneous platforms.Active service is a newly on-demand computing pattern based on the evolution of Web service. It provides users with an integrated, intelligent and personalized Web services solutions. According to users'service requirements, search, discover and dig out the components that meet requirements of users from local network or Internet, and then assemble, compile and publish the service to users. The greatest feature of active service is customization.As the heterogeneity of components on Internet, active service proposes UCDL, which is short for Universal Component Description Language. UCDL is lack of semantic information, which causes some difficulties to search components when the users don't have the relevant knowledge to supply the accurate requirements. So the thesis proposes OCDL to describe component.After the proposition of semantic component description, components in local component resource library are constructed by OCDL, and they are organized and managed uniformly. In order to improve the recall rate and precision rate while making the component retrieval, the thesis proposes the ontology knowledge library based on the LCRW. The ontology knowledge library consists of domain ontology and component ontology, and it represents the semantic layer.As lots of heterogeneous components on Internet, there are many candidate components with the same function retrieved by the retrieval system. It is difficult to choose component from the candidates to assemble the service the next step. The thesis proposes to rank the candidate components by the matching degree and gives the rank algorithm, which will greatly enhance the efficiency of component assembly. The matching degree is calculated by the function matching degree which is based on VSM and parameter matching degree. By the analysis of the final experiment, the predefined component ranking can be replaced by the rank calculated by the algorithm, which can accurately reflected the closer degree between candidate components and user requirements.