Ontology-based Data Source Mapping Method And Strategy

Nowadays, following with the development of network technology and the popularization of network applications, researchers gradually pay more attention to the Web databases whose data are richer and have higher-quality than the traditional static HTML Web pages. However, in some specific research areas, due to difference of designers' understandings about the same database schemas, the data models which describe the same database will be different in the levels of grammar, structure and semantics. Within all these differences, the semantic heterogeneity is the most important one because it will directly influence the application effect of the Web data integration system. Thus, How to eliminate the semantic heterogeneity becomes the core in the Web data integration field.In order to solve the semantic heterogeneity problems, the ontology is introduced in this paper. Ontology is a useful tool for describing the knowledge in specific domain. It supplies a set of concepts and relationships among them which can be used for the whole system, and simultaneously, it provides the evidence and classification method for the database schemas waiting for mapping in domain level, so it can be well used to solve the semantic heterogeneity problems in web data integration system.This thesis briefly introduced the researching status and mapping tools of ontology and data sources mapping method in Deep Web field. On this basis, the paper mainly focused on the building methods and process of the mappings. This paper has given a detailed introduction from the beginning of pre-treatment, including the creation of domain taxonomy, taxonomy ontology and the annotation of the data sources, to the final step of creating mappings and the discussion about searching methods. At last, there are three different searching methods are presented in this paper, and respectively, each one has been illustrated by an example. Through the introduction of this paper, the semantic heterogeneity in Web data integration systems will be solved effectively and also, it can give a variety of answers to users'query through different searching methods.