Empirical Validation of AI-Driven Automated Generation for Privacy Requirements Reuse (DATASET)

This Zenodo record provides the replication package for an empirical validation of AI-driven automated generation for privacy requirements reuse. The package includes (i) privacy requirements catalogs automatically generated with RequiCreator, (ii) equivalent catalogs authored by human experts as a baseline, and (iii) the complete validation outputs (per-catalog results and SPSS Statistics projects) organised across four quality dimensions: Ambiguity, Consistency, Readability, and Redundancy. The validation workflow is structured by branch, each containing executable outputs (results/) and the corresponding statistical artefacts (spss_statistics/). Additionally, the Consistency branch includes traceability outputs produced by RequiTrace (linking requirements to their normative sources) as part of the consistency/traceability analysis. Package contents (high-level): ai_catalogs/: RequiCreator-generated JSON privacy requirements catalogs. human_catalogs/: Human expert-authored JSON catalogs (baseline). validation/: Branch-wise validation outputs for ambiguity, consistency (incl. traceability), readability, and redundancy; includes per-run results and SPSS projects.