Communications of the ACM
A Lightweight Methodology For Rapid Ontology Engineering
Credit: iStockPhoto.com
We are living in a reality that, thanks to economic globalization and the Internet, is increasingly interconnected and complex. There is thus a growing need for semantic technology solutions that can help us better understand it, particularly from a conceptual point of view. Ontologies represent an essential component to developing the Web of Data (such as Linked Open Data1) and Semantic Web applications. An ontology is a conceptual model of (a fragment of) an observed reality; it is, in essence, a repository of interlinked concepts pertaining to a given application domain. Traditionally, construction of an ontology (and its constant evolution, necessary to keep it aligned with reality) is lengthy and costly.
Key Insights
A high-quality ontology requires a rigorous, systematic engineering approach. Existing methodologies are quite complex, conceived primarily for skilled ontology engineers trained to develop large, industrial-strength ontologies. However, before embarking on a full-scale ontology project, it is useful to pursue pilot projects with experimental implementations, testing the applicability of semantic technologies in a confined enterprise area. From this perspective, available ontology engineering methodologies are often unsuitable, overly complex, and demanding in terms of time, cost, and skilled human resources.
No entries found