Evaluating Ontology Cleaning

Christopher A. Welty, Ruchi Mahindru, Jennifer Chu-Carroll

Abstract: Ontology as a discipline of Computer Science has made many claims about its usefulness, however to date there has been very little evaluation of those claims. We present the results of an experiment using a hybrid search system with a significant knowledge-based component to measure, using precision and recall, the impact of improving the quality of an ontology on overall performance. We demonstrate that improving the ontology using OntoClean (Guarino and Welty, 2002)., does positively impact performance, and that having knowledge of the search domain is more effective than domain-knowledge-free search techniques such as link analysis.