An Ontology for Classifying the Semantics of Relationships in Database Design

Relationships are an integral part of conceptual database design because they represent associations between entities from the real world. With the proliferation of both traditional, corporate, and now, web-based, databases representing similar entities, comparison of relationships across these databases is increasingly important. The ability to compare and resolve similarities in relationships in an automated manner is critical for merging diverse database designs, thus facilitating their effective use and reuse. The key to automating such comparisons is the capture and classification of the semantics of the relationship verb phrases. This research presents a multi-layered ontology for classifying verb phrases to capture their semantics. The fundamental layer captures the most elementary relationships between entities. The next, generic layer combines or arranges fundamental relationships to represent generic, realworld relationships. Finally, the contextual layer instantiates the generic relationships with specific verb phrases in different situations. Empirical testing of the ontology illustrates its effectiveness.