Research And Implementation Of Ontology Storage Based On Relational Database

AS the extension of current WWW(World Wild Web), the target of semantic Webis to provide a universal semantic framework for the realization of data sharing andintegrating among different applications. Ontology is the basis of semantic Web andrelational database is the storing way of most data on WWW. So, mutual mappingbetween ontology and relational database scheme is an important part of research onsemantic web, research on ontology storing based on relational database is one ofrelevant main issues.Firstly, this dissertation introduced situation of ontology storing research andanalyzed the issues existing in this field in details. Then, the relevant technologies, suchas ontology describing language and ontology storing pattern, etc. are outlined andcompared with each other from some aspects related with ontology storing.The researching works of this dissertation are as follows:①On the basis of comparing the pros and cons of current ontology storing patternsbased on relational database, the dissertation summarized the general principles ofstoring ontology in relational database.②According to the summarized general principles of storing ontology in relationaldatabase, this dissertation designed and implemented a new mixed storing pattern calledNROSP(Rdb_based Ontology Storing Pattern) of clear structure, high queryingefficiency, complete semantic information and good extension performance.③The dissertation studied the optimization of ontology storing with maturerelational database technology on the basis of NROSP. On the one hand, the importanceof materialized view, clustered index and stored procedure in the optimization of queryefficiency are analyzed. On the other hand, this dissertation summarized the principlesof updating automatically ontology information with reference integrity constraint andupdating trigger under different situations.④The dissertation designed and implemented a ontology storing knowledgemanagement platform system on the basis of optimized NROSP.At the same time, a group of reasonable experiment solutions testing contrastivelyquerying efficiency and extension performance are designed, and vertical pattern andmixed pattern NROSP which is proposed in this dissertation are compared. The resultshows that NROSP has better querying efficiency and better extension performance, and NROSP is more suitable for high efficiency storing of large scale ontology.