Generating OWL ontology from relational database

Abstract

The Semantic Web aims to enrich the existing Web by building an infrastructure layer of a web of data. Three-quarters of the data on the Web come from relational databases. Thus, in the context of the Semantic Web, it has become one of the most important topics to generate OWL ontologies from the legacy databases. For this purpose, the existing approaches typically define how to map the databases to the ontologies. However, the mappings are designed to work under highly normalized database schemas and to be carried out through a batch process of dumping database contents to the ontology repository for ontology population. The approaches can't contribute usefully to the proliferation of ontologies. In this paper, we propose a new set of mapping rules based on a view-based approach, which enables ontologies to be served even without their population through the massive dump process in advance. Consequently, the ontologies out of our approach can play as the complete OWL-described view on the database because they explicitly represent the semantics of relational model objects and operations. Compared with existing approaches, our approach is applicable to a wider range of database schemas, and besides it is designed to extract more meaningful relationships between entities by reasoning. © 2012 IEEE.