A Novel Approach for Addressing Scalability and Computational Complexity Issues of the Ontology Integration Process

The knowledge representation approach plays a vital role in today's era. This enables ontology engineers to form a knowledge base and share it across artificial intelligence (AI) based applications. AI-based applications now use ontologies. This could result in the formation of semantically similar heterogeneous ontologies. This research focuses on the integration of heterogeneous ontologies with the help of matching and merging processes to improve knowledge representation. The existing work focuses on the integration process of low-scale to middle-scale heterogeneous ontologies. This research focuses primarily on the integration of large-scale ontologies. To support the integration of large-scale ontologies, the scalability factor is important. Therefore, the methodologies identified by this proposed work involve preprocessing the input ontologies and creating RDF graphs. These RDF graphs are divided into small-scale subgraphs using novel partitioning algorithms. The matching process identifies the similarity among the ontologies via the large language model. The heavy load of the matching process is evenly balanced through the load-balancing approach during parallel computing. The load balancing approach is designed via the ant colony optimization algorithm. The performance of the proposed model is tested on the BioML, Human Phenotype and Disease benchmark ontology datasets. This results in a reduction in the computational time. The qualitative analysis shows that the integrated ontology is consistent.