The Music Ontology provides a framework for publishing structured music-related data on the Web, ranging from editorial data to temporal annotations of audio signals. It has been used extensively, for example in the DBTune project and on the BBC Music website. Until now it hasn't been systematically evaluated and compared to other frameworks for handling music-related data. In this article, we design a 'query-driven' ontology evaluation framework capturing the intended use of this ontology. We aggregate a large set of real-world music-related user needs, and evaluate how much of it is expressible within our ontological framework. This gives us a quantitative measure of how well our ontology could support a system addressing these real-world user needs. We also provide some statistical insights in terms of lexical coverage for comparison with related description frameworks and identify areas within the ontology that could be improved.
展开▼
