N-Dimensional Matrix-Based Ontology: A Novel Model to Represent Ontologies

This article describes how the assessment of semantic similarities between word pairs is an important component of understanding text which enables processing, classifying and structuring of textual resources. For this purpose, an ontology is a powerful technique when applied to compute similarity. In this article, the authors propose a novel model to represent an ontology in which an N-dimensional matrix is applied, called an N-dimensional matrix-based ontology. This matrix-based ontology attempts to decrease the time complexity of computation. Second, a new semantic similarity measure is introduced and is performed on the N-dimensional matrix-based ontology. Third, the validation of the result of the N-dimensional matrix-based ontology is compared with related studies comparing two well-known benchmarks. The results reveal that in an N-dimensional matrix-based ontology with increasing N, the accuracy of the proposed semantic similarity measure is increased. Moreover, a matrix-based ontology decreases the time complexity when compared to a graph-based ontology.