Learning Class Disjointness Axioms Using Grammatical Evolution

International audienceoday, with the development of the Semantic Web, LinkedOpen Data (LOD), expressed using the Resource Description Frame-work (RDF), has reached the status of “big data” and can be consideredas a giant data resource from which knowledge can be discovered. Theprocess of learning knowledge defined in terms of OWL 2 axioms fromthe RDF datasets can be viewed as a special case of knowledge discov-ery from data or “data mining”, which can be called “RDF mining”.The approaches to automated generation of the axioms from recordedRDF facts on the Web may be regarded as a case of inductive reasoningand ontology learning. The instances, represented by RDF triples, playthe role of specific observations, from which axioms can be extracted bygeneralization. Based on the insight that discovering new knowledge isessentially an evolutionary process, whereby hypotheses are generatedby some heuristic mechanism and then tested against the available evi-dence, so that only the best hypotheses survive, we propose the use ofGrammatical Evolution, one type of evolutionary algorithm, for miningdisjointness OWL 2 axioms from an RDF data repository such as DBpe-dia. For the evaluation of candidate axioms against the DBpedia dataset,we adopt an approach based on possibility theory