LEARNING ONTOLOGY FROM RELATIONAL DATABASE:AN AUTOMATIC ONTOLOGY CONSTRUCTION APPROACH

摘要

Ontology is one of the ways to represent domain specificknowledge and recommended standard such as OWL helpsto share the domain specific knowledge. However, atpresent most of the ontologies are build manually bydomain experts and it needs lots of laborious tedious work,ineffective and error-prone. In this paper we present apractical approach of automatic OWL ontology constructionfrom relational database. The proposed approach willgenerate classes, subclasses, object properties, datatypeproperties, and cardinality restrictions from relationaldatabase schema automatically. First, we build domainconcepts; learn concept relations, properties, andrestrictions. Then, compare ontology created fromnormalized (3NF) and un-normalized (1NF) RelationalDatabases.