Building Local Ontology from Database Relations in Data Integration

Abstract

Source models help in integrating distributed data sources. They need to be represented in conceptual model in order to provide scalable and flexible data integration of autonomous and heterogeneous data sources. Ontology is explicit specification of conceptualization and can handle semantic heterogeneity of the local sources. The technique in this paper, maps database relations to OWL based ontology for the local sources. It minimizes errors and effort involved in manual ontology building. Its distinguished feature is to create ontology from relations in the absence of necessary metadata. Compared with existing techniques, the technique builds ontology automatically and allows adding additional semantics. The technique is presented with examples in this paper.