An evaluation of agile Ontology Engineering Methodologies for the digital transformation of companies

Abstract

Ontologies are increasingly recognised among the key enablers of the digital transformation of knowledge management processes, but still with a low level of adoption in manufacturing companies. Because ontologies and underlying technologies are complex, Ontology Engineering Methodologies (OEMs) provide a set of guidelines to move from an informal to a formal representation of the company’s knowledge base. This study evaluates three agile OEMs, i.e. UPONLite, SAMOD and RapidOWL, in terms of their process and outcome features, i.e. the OEM steps and the expected quality of the ontological models produced. The assessment is performed from the viewpoint of developers of ontology-based technologies in real industrial use cases. Results show that the three agile OEMs reflect different features to effectively support the digital transformation of companies' knowledge management; thus, they cannot be interchangeable. UPONLite is more effective in contexts where there is a lack of skills in OE, with the need for a structured approach in involving domain experts and generating documentation. SAMOD requires a more extended development period, but with several cycles that allow to map different types of knowledge and enable a “try-and-learn” approach. Conversely, RapidOWL lacks a structured sequence of modelling activities and encourages developers to be creative, but at the same time requires higher expertise in OE. Thus, companies and personnel dedicated to OE should choose the methodology according to the main aims guiding their digitalisation process, the current development status, and the level of expertise.