Extracting and Merging Contextualized Ontology Modules

Abstract

Ontology module extraction, from a large ontology, leads to the generation of a specialized knowledge model that is pertinent to specific problems. Existing ontology module extraction methods tend to either render a too generalized or a too restricted ontology module that at times does not capture the entire semantics of the source ontology. We present an ontology module extraction method that extracts a contextualized ontology module whilst extending the semantics of the extracted concepts and their relationships in the ontology module. Our approach features the following tenets (i) identifying the user-selected concepts that are pertinent for the problem-context at hand; (ii) extracting the user-selected concepts, their roles and their individuals; and (iii) extracting other concepts, roles and individuals that are structurally-connected with the user-selected concepts. We apply our ontology module extraction method in the Healthcare domain, and demonstrate (a) extraction of ontology modules from three prostate cancer pathway ontologies; and then (b) merging of extracted ontology modules to generate a comprehensive therapeutic work-flow knowledge for prostate cancer care management.