Pattern-based Interoperability Framework from OWL to Conceptual Modelling Languages

One of the primary distinguishing features of any conceptual model lies in its association with initial conceptual space, often referencing part or all of an ontology provided by the community of practice. Typically, this ontology is specified in OWL to foster reusability through the web. However, a definite mapping from OWL 2 to popular conceptual modelling languages like ORM, UML, or EER that helps in this process is missing. This work addresses this interoperability problem by proposing a tool-based framework to automatically generate conceptual model patterns through querying OWL reasoners. Intending to closely approximate the generated model to the original OWL 2 ontology, we identify five axiom patterns and their representations in ORM 2. Subsequently, we develop a procedure to import partial knowledge from ontologies by exhaustively searching the inferred axioms, normalising them, and translating them. We assess the feasibility of our approach by importing a set of publicly available ontologies and measuring the generated conceptual models using structural metrics. As a result of this experiment, we successfully imported around 40% of the LOV repository ontologies, with around 80% of the axioms recognised and translated.