An extension to the data-driven ontology evaluation

Abstract

Within the semantic web domain, ontologies are an important artifact. Such words as "pivotal" have been associated with the role they play on the semantic web. The role they play on the semantic web as well as their potential for reuse and the proliferation of ontologies in existence have heightened the need for their evaluation. They have been seen as approximate representations of the domain, thus their evaluation concerns itself with the degree of their approximation. This research deemed domain knowledge on which data-driven ontology evaluation is based to be dynamic. This is contrary to the underlying assumptions of current research in data-driven ontology evaluation. The paper hence proposes a multidimensional view to data-driven ontology evaluation that accounts for bias in the valuation of ontologies. The direct contribution to the body of knowledge is a theoretical framework that exposes these biases.