Facilitating Ontological Benchmark Construction Using Similarity Computation and Formal Concept Analysis

ABSTRACT The demo paper deals with preparation of ontological benchmarks as ontology sets selected from ontology repositories. We distinguish between two types of ontology sets according to coverage of ontology metrics values: a homogeneous set contains ontologies with very similar values of selected ontology metrics, while a heterogeneous set contains ontologies with different values of such metrics. Homogeneous sets can be built straightforwardly upon specification of the required metrics values by the user when interacting with repositories. However, building heterogeneous ontology sets is not covered in the current functionality of ontology repositories. We propose to employ search techniques leveraging on ontology similarity in this process. In similarity-based search one can obtain either homogeneous or heterogeneous sets, by varying the selection parameters. Further, we investigate whether Formal Concept Analysis could help generate sufficiently large homogeneous sets