Business process driven ontology discovery method from distributed data environment

Ontology is considered to be one formal description of domain knowledge which is often defined by experts manually. Ontology technique plays an important role in data integrating and semantic web searching. However, it is a hard work for experts to abstract ontology network for complex application, especially from the heterogeneous data sources. In this paper, automatic ontology discovery method is researched to generate local ontology network from relational data base schema. The data column characteristic, primary key and foreign key relations of the separated data base dictionary are analyzed. Then, ontology reasoning is used to eliminate the inconsistence among the distributed ontology networks. Business process driven method is proposed to merge the local ontology network into a global ontology network to reflect the enterprise knowledge. Finally, the method is demonstrated in a hospital ontology system constructing. The result shows that ontology discovery method could abstract initial ontology efficiently from data base and the local ontology could be combined to enterprise global ontology.