Antipattern detection: How to debug an ontology without a reasoner
Résumé
In ontology design, an Ontology Design Pattern (ODP) is a modeling solution to a recurrent ontology design problems. As opposed to ODP, antipatterns are bad modeling practices. This paper deals with detection of antipattern for debugging purpose of huge ontologies. We focus on the detection of the Onlyness Is Loneliness (OIL) antipattern. We propose an antipattern detection method based on ontology transformations and SPARQL queries. This approach does not need reasoner to detect antipattern. Our method detects candidates of OIL antipattern. These candidates localize class definitions where OIL occurrences can appear. This enables to draw ontology developer's attention to avoid errors during ontology development process. We conduct some experiments to detect OIL antipattern in an OWL ontology corpus obtained from the Watson ontology search engine.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...