Ontology Authoring from Relational Database: A Model Based Approach

The ontology authoring is a fundamental task in the Semantic Web. This process enables the domain expert to develop ontologies with the help of dedicated tools. This article presents an approach for building OWL ontologies from relational databases based on Model Driven Engineering (MDE). The proposed approach consists of two phases: (1) Preprocessing phase and (2) Transformation phase. The first one consists of creating an input model from the database which must conform to its meta-model. The second phase takes this model as input and transforms it into an OWL file by executing a set of mapping rules written in Atlas Transformation Language (ATL). The transformation process is done at a higher level of abstraction; it does a matching between the source meta-model elements (Database) and the target meta-model elements (OWL). We have concretized our approach as the DB2OWLOntology tool, and we have evaluated it with a set of databases. The obtained results are encouraging and show the efficiency of the proposed approach.