Translating relational & Object-Relational Database models into OWL models

We describe an extensible framework for translating data models into ontology models. Initially, the framework addresses two types of source data models: the relational database (RDB) and object-relational database (ORDB) models. The derived ontology model is based on OWL. The framework extracts information about the source data models from the metadata maintained by the DBMS. The extracted metadata includes most of the integrity constraints that are typically maintained by a DBMS. The extracted metadata is then analyzed to identify ontology concepts, properties, and explicit relationships, discover redundant ontology concepts and implicit relationships, and identify restrictions on properties and relationships. The analysis is based on heuristic database modeling techniques. The analyzed data is automatically translated into a rudimentary OWL ontology model that can be enhanced by ontology modelers. The paper provides examples to demonstrate how the translation is conducted.