Title: Large-scale ontologies: pattern and partition-based alignment

Authors: Soumaya Kasri; Fouzia Benchikha

Addresses: Department of Computer Science, University of August, 20 1955 – Skikda road el hadaiek, BP 26 Skikda 21000, Algeria. ' LIRE Laboratory, Mentouri University of Constantine, road Ain el Bey, 25000 Constantine, Algeria

Abstract: Several ontologies covering the same domain or overlapping domains have been independently built of each others. They are often a source of heterogeneity. In order to make them interoperable, the semantic links between their entities must be established. This task is called |ontologies alignment|. However, the proposed alignment methods meet the challenge of |scaling up|. To take up this challenge, we propose an alignment method with partitioning. This method consists of partitioning each ontology into blocks around the anchors. In the alignment step, each block of the first ontology is aligned with one block of the second one. The results obtained in the evaluation of our method on selected ontologies show its efficiency.

Keywords: large scale ontologies; ontology partitioning; clustering algorithms; ontology design patterns; ontology alignment; web science.

DOI: 10.1504/IJWS.2011.044081

International Journal of Web Science, 2011 Vol.1 No.1/2, pp.36 - 53

Published online: 28 Mar 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article