A Method for Ontology Alignment Based on Semantics of Attributes

Abstract

In this article, a method for ontology alignment, which is based on using semantics of attributes describing concepts, is presented. This work is an extended version of our previous works on ontology alignment, which is the task of designating correspondences between two ontologies. After investigating recent approaches, we noticed a lack of analysis of the lowest level of expressing knowledge within them. We treat attributes of concepts as this level and we claim that embedding their precise definition, semantics, and the ways in which they can interact with each other in the process of mapping ontologies can enhance former solutions of this problem. We show that developing such an approach brings us closer to a consistent and more intuitive methodology of aligning ontologies.

Working framework of semantic interoperability for CRIS with heterogeneous data sources

“…During the process of bibliographic revision a number of evaluation methodologies were identified, such as the proposal by Dellschaft, mainly focused on the quality of the ontologies (Dellschaft and Staab, 2006); or those that focus on evaluating the alignment of ontologies according to the indexes of recall and precision ( Jérôme and Euzenat, 2008), with operational weaknesses in the context of interoperability, owing to the fact thatin our opinionthey should be applied after the evaluation of the technique of interoperability, as the methodology has more to do with information retrieval. Other works propose the evaluation of aligned vocabularies using recall and precision in the process of retrieval of certain terms of the benchmarking domain (Pietranik and Ngoc-Thanh, 2012). The studies by Gal, in turn, look only into the evaluation of the Big Data processes (Gal, 2014).…”

“…This approach was used in further studies with noteworthy variations of the logical modeling and context of application (Wajid and Marín, 2009). Other solutions for aligning and transforming ontologies address the issue from an approach focused on knowledge representation, paying more attention to the attributes than to the underlying logical structure and the connections between contexts, so as to avoid problems of inconsistency between the classes of ontologies and the structures they define (Pietranik and Ngoc-Thanh, 2012). Most semantic alignment solutions insist on the calculation of similarity by means of different methods, among which are:…”

Matching Large Scale Ontologies Based on Filter and Verification

“…e larger the scale of ontology, the greater the heterogeneity between ontologies and the more difficult to utilize structural information to calculate structural similarities between entities [11]. As mentioned before, 50% alignments were obtained by calculating structural similarities between entities.…”