A novel method for improving the efficiency of automatic construction of ontology from a relational database

Abstract

With the development of the semantic web, ontology is playing an increasingly important role in many research areas such as semantic interoperability and knowledge base. However, constructing ontology manually is complicated and needs the supports of domain experts in knowledge acquisition as well, so it is time-consuming, error-prone and tedious-work. Learning ontology from existing resources is a good solution. We can use relational database for building ontology, because relational database is widely used for storing data. This paper proposes an approach of learning ontology web language (OWL) from data in relational database. Compared with existing methods, our approach can acquire ontology from relational database automatically. In addition, our proposed method, unlike other existing methods, all types of relationships between tables are considered. The proposed method is implemented using Jena and MySQL and is applied on a sample relational database (RDB). The resulting ontology was shown as an OWL file. The evaluation of the generated ontology will use FaCT + + and Pellet. Key words: Ontology building, OWL, Relational database, Semantic web, mapping rules.