Ontological ISA-Distance Measure for Information Visualisation on Conceptual Maps
Résumé
Finding the right semantic distance to be used for information research, classification or text clustering using Natural Language Processing is a problem studied in several domains of computer science. We focus on measurements that are real distances: i.e. that satisfy all the properties of a distance. This paper presents one ISA-distance measurement that may be applied to taxonomies. This distance, combined with a distance based on relations other than ISA, may be a step towards a real semantic distance for ontologies. After presenting the purpose of this work and the position of our approach within the literature, we formally detail our ISA-distance. It is extended to other relations and used to obtain a MDS projection of a musical ontology in an industrial project. The utility of such a distance in visualization, navigation, information research and ontology engineering is underlined.