Ontology Development Approach Adopting Analogy and Competency Questions

Ontologies provide a useful format for structuring arbitrary knowledge domains. However, their development is a time-consuming and costly process, that requires not only domain expertise, but also knowledge and skills for mapping the domain knowledge to the ontology language of classes, instances and relations. In this paper, we propose a novel method for developing ontologies that exploits analogies between different but related knowledge domains represented by ontologies. Our approach processes two ontologies for related problems or knowledge domains, where one of them is complete, while another needs further development; it then generates a series of competency questions for the domain expert in order to extend the new ontology which will be generated as outcome. The proposed method is evaluated in case of ontology from wireless networking domain – creation of ZigBee ontology starting from the one for WiFi. The main implications of our method are twofold: 1) reduces the time need for an ontology engineer from the ontology development process, and it reuses the domain expertise already encoded in the original ontology making the process faster and less costly 2) provides further cognitive links between different knowledge domain. Our approach also opens some new directions in the knowledge discovery and management research regarding the role of machine-learning algorithms in ontology engineering.