Should ontologies be evaluated?

Published: September 01, 2016

Author(s)

Judith R. Gelernter, Jyoti Jha

Abstract

Ontologies are semantic tools for Artificial Intelligence. Ontologies list and define objects in a domain and the relationship among objects such as to do reasoning within and among systems. Mistakes within ontologies bring system errors if their use in a system is rigid. But we have found that ontology reasoners cannot find a range of errors. Therefore, with present tools, we suggest that ontology evaluation is not worthwhile. Instead, we must re-make current tools so that it is unnecessary to know ontology languages or logic for domain experts to edit an ontology, or else we could adjust rules in ontology-based systems to allow fuzzy matching based on ontologies that are imperfect.
Citation: Journal of Data and Information Quality
Volume: 7
Issue: 3
Pub Type: Others

Keywords

Ontology evaluation, ontology reliability, accuracy, error correction
Created September 01, 2016, Updated November 10, 2018