A multi-layered ontology for comparing relationship semantics in conceptual models of databases

Article type: Research Article

Authors: Purao, Sandeep | Storey, Veda C.

Affiliations: School of Information Sciences and Technology, The Pennsylvania State University, PA, USA. E-mail: [email protected] | Department of Computers and Information Systems, Georgia State University, Atlanta, GA, USA. E-mail: [email protected]

Abstract: Relationships are an integral part of the design of a database. Comparing and integrating relationships from heterogeneous databases requires that the relationships be mapped to each other or to a common classification. Identifying similarities and resolving differences in relationships across large data sources is a resource-intensive task that could benefit greatly from semi-automated approaches. A prerequisite to developing such approaches is a clear understanding of the semantics of relationships used in database design. This research presents a layered ontology for classifying the semantics of relationships. It consists of a core layer that captures the fundamental types of relationships between entities. A middle layer provides the internal context, obtained from entities surrounding the relationship, to interpret the fundamental types. The outer layer allows further interpretation using the external context, that is, the domain in which a relationship is being used. An initial assessment on relationships from a variety of application domains demonstrates that the ontology can be adequate and useful for comparing relationships across databases.

Keywords: Database design, relationships, verb phrase, ontology, semiotics, entity-relationship model, natural language, conceptual modeling, classification scheme

Journal: Applied Ontology, vol. 1, no. 1, pp. 117-139, 2005