Authors: Warith Eddine Djeddi; Mohamed Tarek Khadir
Addresses: Laboratoire sur la Gestion Electronique de Document (LabGED), Computer Science Department, University of Badji Mokhtar, P.O. Box 12, 23000 Annaba, Algeria ' Laboratoire sur la Gestion Electronique de Document (LabGED), Computer Science Department, University of Badji Mokhtar, P.O. Box 12, 23000 Annaba, Algeria
Abstract: Achieving high match accuracy for a large variety of ontologies, considering a single matcher is often not sufficient for high match quality. Therefore, combining the corresponding weights for different semantic aspects, reflecting their different importance (or contributions) becomes unavoidable for ontology matching. Combining multiple measures into a single similarity metric has been traditionally solved using weights determined manually by an expert, or calculated through general methods (e.g. average or sigmoid function), however this does not provide a flexible and self-configuring matching tool. In this paper, an intelligent combination using Artificial Neural Network (ANN) as a machine learning-based method to ascertain how to combine multiple similarity measures into a single aggregated metric with the final aim of improving the ontology alignment quality is proposed. XMap++ is applied to benchmark and anatomy tests at OAEI campaign 2012. Results show that neural network boosts the performance in most cases, and that the proposed novel approach is competitive with top-ranked system.
Keywords: ontology alignment; context measures; WordNet; ANNs; artificial neural networks; learning; overfitting; cross-validation; large-scale ontologies; similarity measures.
International Journal of Metadata, Semantics and Ontologies, 2013 Vol.8 No.1, pp.75 - 92
Received: 01 Oct 2012
Accepted: 21 Feb 2013
Published online: 28 May 2013 *