Automatic Ontology Learning from Textual Documents: A Systematic Literature Review

Abstract

This paper aims to address the fundamental processes involved in creating an automatic ontology from a specific domain text, by employing a comprehensive research framework. Other existing SLRs were not so specific on the architecture used for ontology learning. The document answers four research questions related to the efficiency of artificial intelligence techniques in automatic ontology creation. The methodology involves a systematic literature review using the PRISMA framework, resulting in the selection of relevant articles. The study highlights the role of similitude between ontologies as performance metric. Finally, human supervision in ontology learning algorithms is still required due to errors introduced by unsupervised approaches was one of most important findings.