Framework for Enhanced Ontology Alignment using BERT-Based

Abstract

This framework combines a few approaches to improve ontology alignment by using the data mining method with BERT. The method utilizes data mining techniques to identify the optimal characteristics for picking the data attributes of instances to match ontologies. Furthermore, this framework was developed to improve current precision and recall measures for ontology matching techniques. Since knowledge integration began, the main requirement for ontology alignment has always been syntactic and structural matching. This article presents a new approach that employs advanced methods like data mining and BERT embeddings to produce more expansive and contextually aware ontology alignment. The proposed system exploits contextual representation of BERT, semantic understanding, feature extraction, and pattern recognition through data mining techniques. The objective is to combine data-driven insights with semantic representation advantages to enhance accuracy and efficiency in the ontology alignment process. The evaluation conducted using annotated datasets as well as traditional approaches demonstrates how effective and adaptable, according to domains, our proposed framework is across several domains.