Research On Ontology Construction Method Based On Relational Database

Along with the rapid development of Internet and information technology, network functions are also unceasingly changing. From merely online data publishing, to a platform for information exchange, which further evolve to today's intelligent search. However, because of professional field information independency and the limitations of current search engines, the query results are unsatisfactory,the returned information usually cannot meet the real needs of use. Thus, the founder of web, Tim Berners-Lee proposed the Semantic Web to solve this problem. Semantic Web utilizes multi-level expression framework, in which ontology is an important component.Ontology was previously a concept in philosophy research area. It was later applied to other fields. Ontology is a kind of model used to describe the knowledge system in the semantic and knowledge layer, which can effective solve the problem of knowledge sharing and knowledge reuse. Therefore, domestic and foreign researchers are increasingly paying concern to Ontology, and it is used widely in many areas, such as knowledge engineering, information searching and semantic web.As a fundamental work of ontology study, ontology construction has attracted many domestic and foreign researchers. Nowadays, most of the ontology construction is carried out by hand. However, building ontologies manually is a time-consuming work. Although there are many ontology editing tools offering ontology editing capabilities, people must obtain the domain concepts and their relations artificially, which requires the participation of experts in the field. When there are many concepts in the field, ontology construction is a heavy amount of work. Since the ontology construction has become a bottleneck that prevents its application, the applications based on ontologies are very difficult to advance.Under this background, in order to make the ontology construction more simpler and more efficient in order to reduce the heavy workload,domestic and foreign scholars begin to look into automatic ontology construction. Many scholars consider using the existing resources and technology to build ontology. As the database technology gradually become mature and network becomes omnipresent rapidly, database has become a kind of important tool, which many professions and section uses to store large amount of data. Relational database, which is most widely used nowdays, has the following characteristics. First, most of the relational databases are constructed towards specific application domains, resulting in high correlations in domains. Second, in the relational database the information format is neat, so it is advantageous for the automated analysis and extraction. Third, relational database model is obtained through a large number of need analysis of areas by the developers in the initial period, which in fact conceals the domain conceptual model. Hence, relational database is good source of ontology construction.In this paper, through the research of semi-automatic ontology construction, we try to find a practicable method of semi-automatic ontology construction, which improves ontology construction in both efficiency and quality, while reducing the artificial workload of the construction process.Focusing on the relational database features, this paper proposes a kind of ontology construction method based on relational database. The details are as follows: First, through a detailed analysis of relational database schema, we define a series of rules that extracts data in a relational database to ontology, and follow these rules to build an preliminary ontology. Then, we use WordNet to expand the initial ontology. Finally, we use ontology editing tools to modify and improve the ontology manually.The research of ontology enables us to gain a more comprehensive understanding of ontology, including ontology definition, ontology classification, ontology language, application of ontology, etc. Fundamental knowledge is a great help to build ontology. In ontology construction, we propose a method of ontology construction based on the relational database and build a system of ontology construction based on relational database. The system mainly includes the following modules: relational database schema extraction module, ontology construction module, WordNet-based ontology expansion module and perfecting module. The system uses the MYSQL database and java JDBC API to extract the relational database schema. In ontology construction module, we use the Jena technology to carry on the automatic ontology construction. In WordNet-based ontology expansion modules, we extend the initial ontology in both hierarchy and synonyms by WordNet. In perfecting modules, Protégésoftware is used for manual ontology editing. Finally, we tested various functional modules in the system, and verified the feasibility of this method.Because of the limitation of time and personal research experience, this research has some inadequacies. First, the system interface is not friendly enough, so it is necessary to further develop a good user interface. Second, the system only uses MySQL database and does not yet support other relational databases. Finally, in the ontology construction process, the data stored in the database has not yet been fully utilized.