The Research Of Building Domain Ontology Semi-Automatically Based On Relational Database

Ontology as a concept model and tools of building models which can describe knowledge system in the semantic and knowledge layer, has called much attention of domestic and foreign researchers since it was bought forward, and it is used widely in many fields of computer, such as knowledge engineering, digital library, software reusing, information searching, heterogeneous information processing and semantic web.At present most data are stored in relational databases, the information is called Deep Web (Invisible Web). Comparing with Surface Web (Visible Web) composed by static pages, Deep Web can dynamically responses specific requests, so traditional browser can't get these deep information. Contents in Deep Web are richer, more specialized and with higher quality. So how to search and deal with these valuable deep resources automatically is a problem which has been concerned.Semantic heterogeneous information can be eliminated when building ontology, the most usual ones are: name heterogeneous, attributes heterogeneous, equivalent element, crossed element etc. In ontology-based data integration, to build a global ontology and domain ontology which can describe contents of data sources, and define semantic mapping from global ontology to each input model of data source to eliminate semantic heterogeneous problems, thereby to complete semantic query .The purpose of this paper is to build semantic ontology information efficiently based on relational database, which benefits share and mutual operation between scientific databases, this is an efficient way to enhance scientific database from data management to knowledge management. This paper introduces research background at first, analyzes the disadvantages of traditional ontology build technology, put forward the way that build ontology semi-automatically based on relational database; secondly presents related technology of ontology; thirdly discusses some key technologies of building ontology based on relational database, mainly includes the mapping rule definition between relational database and ontology language, domain ontology normalization, according to the contents mentioned above we designs implementing framework, and finally introduces prototype system, forecasts our future efforts.