The field of big data and artificial intelligence is growing at a tremendous rate, generating massive amounts of data and information in the process. It is of utmost importance that methods of managing these large amounts of data and knowledge be developed, such as ontology engineering. Ontologies, however, are highly complex and accordingly ontology engineers emphasize the reuse of existing ontologies as opposed to developing ontologies de novo. Unfortunately, given the vast range of ontologies available online, users are faced with the arduous problem of deciding which ontologies to select for reuse. This study attempts to solve the ontology selection problem by employing a multi-criteria decision-making approach. The ELECTRE II and IV methods are implemented and applied to rank a set of 200 biomedical ontologies obtained from the BioPortal repository. The statistical correlation between the methods is quantified with the use of the Spearman's Rho and Kendall's Tau correlation coefficients. All methods of analysis illustrated that the ELECTRE II and IV methods are suitable for ranking ontologies and depicts comparable results.
DOI: https://doi.org/10.1109/icabcd54961.2022.9856133
Publish Year: 2022
