Research On Relational Model Data Semantic Extraction Method Based On Ontology

Semantic Web is the main direction of development for the existing Internet. In the Semantic Web architecture, ontology is an important component. Because of ontology, Semantic Web can have semantic reasoning functions. Ontology construction is the basic and the first step in application of ontology. The quality of ontology construction has a direct impact on follow-up work. According to the existing methods of ontology construction, most methods are artificial manually edit and require the participation of expert who is in that field. The efficiency and quality of ontology construction are not very good. So, it is very important that how to reduce the participation of experts and improve the efficiency of ontology construction.Firstly, the shortcomings of existing Internet are analyzed in this paper. The basic architecture of Semantic Web is summarized. According to the Semantic Web architecture, the basic concepts of ontology, ontology composition, ontology description language, the rule and the methods of ontology construction are all discussed.Secondly, in the process of ontology construction, there are some different methods for different data ontology construction. So, ontology construction method of the data with structure in particular relational database structure is focused on. According to the existing characteristics of the methods of the relational database semantic information extracted, it is proposed that getting E/R model from the relational database model through reverse engineering. After that, do semantic extraction, which is based on ontology. And the method and rule from E/R model to ontology are given.Finally, according to corresponding rules and methods, using XML analytical tool and Jena framework, the method and the rule from E/R model to ontology is implemented. The semantic information of relation model is saved as an OWL file, which is one of the ontology description languages. Then, the purpose of extracting semantic from relational model is achieved.