Role Assertion Analysis: a proposed method for ontology refinement through assertion learning
Adrien Coulet, Malika Smaïl‐Tabbone, Amedeo Napoli, Marie‐Dominique Devignes · Frontiers in artificial intelligence and applications · 2008
We propose an approach for extending domain knowledge represented in DL ontology by using knowledge extraction methods on ontology assertions. Concept and role assertions are extracted from the ontology in the form of assertion graphs, which are used to generate a formal context manipulated by Formal Concept Analysis methods. The resulting expressions are then represented as DL concepts and roles that can be inserted into the initial ontology after validation by the analyst. We show, through a real-world example, how this approach has been successfully used for discovering new knowledge units in a pharmacogenomics ontology.