Ontology Modularization for Knowledge Selection: Experiments and Evaluations ⋆ Mathieu d’Aquin 1 , Anne Schlicht 2 , Heiner Stuckenschmidt 2 , and Marta Sabou 1 1 Knowledge Media Institute (KMi), The Open University, Milton Keynes, UK {m.daquin, r.m.sabou}@open.ac.uk 2 University of Mannheim, Germany {anne, heiner}@informatik.uni-mannheim.de Abstract. Problems with large monolithical ontologies in terms of reusability, scalability and maintenance have led to an increasing inter- est in modularization techniques for ontologies. Currently, existing work suffers from the fact that the notion of modularization is not as well un- derstood in the context of ontologies as it is in software engineering. In this paper, we experiment on applying state-of-the-art tools for ontology modularization in the context of a concrete application: the automatic selection of knowledge components to be used for Web page annotation and semantic browsing. We conclude that, in a broader context, an eval- uation framework is required to guide the choice of a modularization tool, in accordance with the requirements of the considered application. Keywords: Ontology modularization, partitioning, module extraction 1 Introduction Modularization is a crucial task to allow ontology reuse and exploitation on the Semantic Web. The notion of modularization comes from Software Engineering where it refers to a way of designing software in a clear, well structured way that supports maintenance and reusability. From an ontology engineering perspective, modularization should be considered as a way to structure ontologies, meaning that the construction of a large ontology should be based on the combination of self-contained, independent and reusable knowledge components. In reality, even if they implicitly relate several sub-domains, most of the ontologies are not structured in a modular way. Therefore, in order to facilitate the management and the exploitation of such ontologies, ontology modularization techniques are required to identify and extract significant modules in existing ontologies. While there is a clear need for modularization, there are no well-defined and broadly accepted definitions of modularity for ontologies. Several approaches have been recently proposed to extract modules from ontologies, each of them implementing its own intuition about what a module should contain and what ⋆ This work is partially funded by the Open Knowledge (IST-FF6-027253) and NeOn projects (IST-FF6-027595), and partially supported by the German Science Foun- dation under contract STU 266/1 as part of the Emmy-Noether Program.