PERSPECTIVES OF ONTOLOGY REFACTORING

Abstract

Existing ontologies are supplemented and expanded in scope, increasing their complexity. Over time, integration issues arise with adding new knowledge to an ontology and maintaining existing ones. The problem is exacerbated by the fact that ontologies typically incorporate knowledge from databases and annotated documents in natural language, significantly increasing the volume and complexity of ontologies. This paper focuses on processing and analyzing existing ontologies and applying specific refactoring methods to enhance ontology and improve the readability of presented knowledge for developers and users. The development of corresponding software and metrics is carried out to objectively demonstrate the effectiveness of refactoring methods. During the research process, a certain list of patterns was identified that can be used for ontology refactoring, and one of them was implemented – the "Property Raising" pattern – transferring a property to a superclass. As a result of testing, it was found that after applying this pattern, the volume of the original ontology data file decreased by 3-5%, provided that the subclass contained common properties that were passed to the superclass.