Complex Ontology Alignments using OWL Ontologies Embedding
Houda Akremi, Mouhamed Gaith Ayadi, Sami Zghal · 2023
The semantic heterogeneity issue in the information integration can be solved by using ontology alignment. The goal of the ontology alignment technique is to locate concepts that are semantically identical in two ontologies. But, one of these alignments’ drawbacks is their lack of expressiveness, which can be accounted by using complex alignments. To deal with it, an effective strategy is to model ontologies in vector space and compute their similarity scores to determine the correlation levels. In this paper, a semantic embedding-based ontology matching technique for OWL 2 ontologies is employed to calculate entities’ resemblance in order to increase the alignment accuracy. This technique is reinforced by a stable marriage-based alignment extraction algorithm to establish high-quality matching. The experimental investigation analyses the performance to the benchmark track supplied by the Ontology Alignment Evaluation Initiative (OAEI). Experimental findings demonstrated the effectiveness of our matching method.