Research On Ontology-based Schema Matching

Schema Matching plays a important role in many application domains, such as schema integration,data warehouses,e-commerce,data integration and so on. Because of the variety of the representation of the schema and the complexity of the heterogeneous, at present, the research on schema matching can't find all the corresponding relation between the schema elements. It also can't ensure that all the corresponding relations it finds are right. So the paper does some research on Ontology-based Schema Matching, aiming to assist developing a generic matching system which is with high automatic and can be applied to different application domains.Firstly, the present condition of the current schema matching area is summarized, the importance of combining the ontology with the schema matching is analyzed, and then the method of ontology based schema matching system is proposed in this paper.Secondly we analyze the two important components: similarity finding and similarity combination. Then we give a design of a generic matching library for similarity finding and a method for similarity combination which is automatic tuning and can check the performance of matching system by test suite.Third a matching arithmetic by Latent Semantic Index is represented. And we design a mapping ontology for the represent of matching result.Last we design a matching system prototype and the class diagrams of the main modules are represented. Then serials of experimentation are done and the results are analyzed.