An Approach for Learning Ontology from Relational Database

Abstract

As a conceptual model and modeling tool, ontology can describe knowledge systems at the semantic level, it can also effectively solve the problem of information integration and sharing. In the research of information integration based on ontology, how to transform relational database data into ontology is an important research direction. Since relational databases are widely used to store data, this paper proposes a new method(WN_Graph) for learning ontology from relational database. Compared with the existing methods, WN_Graph combines intermediate conceptual graph model with WordNet to get more hierarchical relationships of concepts. Therefore, more rich semantic relationships of relational databases are extracted. We use data set from the medical field to verify and analyze the proposed method.