Formal Concept Analysis: A Unified Framework for Building and Refining Ontologies

Abstract

Building a domain ontology usually requires several resources of different types, e.g. thesaurus, object taxonomies, terminologies, databases, sets of documents, etc, where objects are described in terms of attributes and relations with other objects. One important and hard problem is to be able to combine and merge knowledge units extracted from these different resources within an homogeneous formal representation (such as a description logic or OWL). The purpose of this article is to show which kinds of resources should be available for designing a real-world ontology in a given application domain, and then how Formal Concept Analysis and its extension - Relational Concept Analysis- can be used for materializing an associated ontology. This resulting target ontology can then be encoded within OWL or a description logic formalism, allowing classification-based reasoning. A real-world example in microbiology is detailed. Finally, an evaluation including tests on recall and precision shows how source resources can be completed with other existing domain resources using a semi-automatic analysis process.