Semantic Modelling and Ontology Integration Dataset

This dataset captures a large-scale, structured compilation of software requirements derived from multiple domains—Non-Functional Requirements (NFR), Healthcare, Automotive, and Financial systems—totaling over 11,800 entries. The first sheet is based on the PROMISE repository's NFR dataset, containing 622 categorized requirements across 11 quality attributes such as performance, security, and usability. Each entry includes confidence scores and keyword annotations from an SREF classifier.

The Healthcare, Automotive, and Financial sheets represent real-world requirements, many extracted or simulated from domain-specific systems. For Healthcare (3456 entries), annotations include type classification, detected issues, and extracted entities. The Automotive sample (4892 entries) integrates ASIL levels and validation status per ISO 26262, while Financial requirements (2847 entries) include priority levels and traceability links.

Supplementary sheets contain a traceability matrix (2145 links) and 699 inconsistencies detected using NLP and ontology tools. A final statistics sheet summarises metrics like average classifier confidence and domain-specific inconsistency counts.

The dataset is ideal for research in requirements classification, traceability, inconsistency detection, and domain adaptation for NLP models in RE.