A Review of Ontology Development Methodologies: The Way Forward for Robust Ontology Design

Abstract

In this present age, the application of ontology as a data modeling technique across different fields of study, for example, knowledge management and information retrieval systems, is indispensable. This development is necessary to find viable solutions to the challenges of data heterogeneity and concept mismatch. Therefore, the end goal is geared toward achieving machine-represented data; in other words, the data are being modeled ontologically. There are existing ontology design methodologies; however, a single methodology is often not complete to design a robust ontology. Thus, this research aims to review the existing standard methodologies through concept-based analysis that suggests a way forward to design robust ontology. The analysis of the review is carried out by considering the goals of achieving robust ontology design, such as data integration, accessibility, reusability, and domain granularity. Based on the literature, this review shows that collaborative design with domain experts, application of standard evaluation techniques, modification of existing ontology development methodologies, types of ontology, and ontology-based machine learning models are determinant factors that define the robustness of ontology. Therefore, if an ontology developer pays attention to these criteria to design an implementable model, this would pave way for robust ontology to be designed.