A Systematic Literature Review on Combining Ontology with Bayesian Network to Support Logical and Probabilistic Reasoning

Abstract

Reasoning in ontology is currently limited to logical reasoning. It is because ontology does not have a standard for probabilistic reasoning. Various approaches have been made by researchers to add the ability for ontological reasoner to do probabilistic reasoning. The approach is done by combining ontology with Bayesian network that does have probabilistic reasoning abilities. This study mapped out various approaches performed in combining ontologies with Bayesian networks to realize logical and probabilistic reasoning simultaneously. We use a systematic literature review method to identify the primary studies on combining ontology with Bayesian network following a predefined review protocol. We searched from four indexing services (SCOPUS, IEEE Xplore, ACM Digital Library, and SpringerLink) and got the result of 74 papers accepted for the review. We extracted properties from these studies and found 8 motivations of the studies, 5 contexts, 4 factors involved, and 5 techniques used in combining ontology with Bayesian network. The aim and the context are clearly stated in most of the studies, while most of the authors did not completely discuss the threats in their papers. As for the method and findings, most of the studies describe and discuss it moderately.