An Alignment-Based Implementation of a Holistic Ontology Integration Method

Inès Osman; Salvatore Flavio Pileggi; Sadok Ben Yahia; Gayo Diallo

doi:10.1016/j.mex.2021.101460

An Alignment-Based Implementation of a Holistic Ontology Integration Method

MethodsX. 2021 Jul 23:8:101460. doi: 10.1016/j.mex.2021.101460. eCollection 2021.

Authors

Inès Osman¹, Salvatore Flavio Pileggi², Sadok Ben Yahia^{1

3}, Gayo Diallo⁴

Affiliations

¹ LIPAH - LR11ES14, Faculty of Sciences of Tunis, University of Tunis El Manar, Tunis 2092, Tunisia.
² School of Information, Systems and Modelling, University of Technology Sydney, Australia.
³ Department of Software Science, Tallinn University of Technology, Estonia.
⁴ INRIA SISTM, Team ERIAS - INSERM Bordeaux Population Health Research Center, University of Bordeaux, F-33000 Bordeaux, France.

Abstract

Despite the intense research activity in the last two decades, ontology integration still presents a number of challenging issues. As ontologies are continuously growing in number, complexity and size and are adopted within open distributed systems such as the Semantic Web, integration becomes a central problem and has to be addressed in a context of increasing scale and heterogeneity. In this paper, we describe a holistic alignment-based method for customized ontology integration. The holistic approach proposes additional challenges as multiple ontologies are jointly integrated at once, in contrast to most common approaches that perform an incremental pairwise ontology integration. By applying consolidated techniques for ontology matching, we investigate the impact on the resulting ontology. The proposed method takes multiple ontologies as well as pairwise alignments and returns a refactored/non-refactored integrated ontology that faithfully preserves the original knowledge of the input ontologies and alignments. We have tested the method on large biomedical ontologies from the LargeBio OAEI track. Results show effectiveness, and overall, a decreased integration cost over multiple ontologies.•OIAR and AROM are two implementations of the proposed method.•OIAR creates a bridge ontology, and AROM creates a fully merged ontology.•The implementation includes the option of ontology refactoring.

Keywords: 00-01; 99-00; Bridge ontology; Full-Merge ontology; Ontology; Ontology integration; Ontology merging; Ontology refactoring; Semantic interoperability.