Conformance testing of ontologies through ontology requirements

Abstract

In recent years, several standard ontologies have been developed to maximise semantic interoperability in different domains; such standard ontologies ensure quality and integrity when describing a domain. Therefore, mechanisms to guarantee that developers build ontologies that conform to such standards are needed. However, while in fields such as Software Engineering or industry, conformance testing plays an essential role during product development, in the Ontology Engineering field there is a lack of techniques for this type of testing. This work introduces an ontology conformance testing method to analyse conformance between an ontology and a standard based on the standard requirements. Grounded on this method, the work also presents a minimum common knowledge identification method for analysing how a group of standards covers a particular domain and for identifying whether there are conflicts between them. This work has been validated by analysing the conformance between an ontology network and a set of standards on the Internet of Things domain, and by analysing the minimum common knowledge between such standards. This analysis shows that the conformance between ontologies and standards is mostly related to definition of classes. Furthermore, the analysis shows that although the analysed standards are related to the same domain, they are created to describe different areas of concern and, thus, there is a minimum overlap between them. Finally, it was concluded that the quality of the conformance analysis depends on the quality of the requirements specification: the more precise the requirements, the more precise the analysis between ontologies and standards.