A Novel Framework for RDF Schema Extraction in NoSQL Databases Using Sentence-BERT

NoSQL databases, known for their flexibility and scalability, have become pivotal in handling diverse and unstructured data. However, their schema-less nature introduces significant challenges in metadata management, query optimization, and data integration. This research presents a novel schema extraction framework that captures the structural and semantic complexity of NoSQL databases. By integrating constraints, logical rules, and Sentence-BERT embeddings, the proposed approach generates semantically enriched schemas that ensure accuracy, coherence, and usability. Experimental evaluation highlights its adaptability across diverse datasets, demonstrating improvements in semantic precision, data integration, and metadata quality. The framework provides a practical solution for bridging schema-less and structured database workflows, enhancing interoperability and analytics capabilities.