CVD XAI multi-paradigm benchmark reproducibility package
Year: 2026
Type: software
Abstract: Executed notebooks, corrected analysis scripts, prediction arrays, model artifacts, figures, and tables for a 15-model cardiovascular disease status classification benchmark with cross-pipeline feature-ranking analysis. Fifteen classifiers were evaluated on one stratified held-out test set (n = 10,281); the corrected stacking ensemble reached ROC-AUC 0.800367 (bootstrap 95% CI 0.791538-0.809095), and no pairwise DeLong comparison among the six highest-ranked models was statistically significant. Licensing is mixed: source code under src/ and tools/ is MIT, while documentation, figures, tables, prediction arrays, and model artifacts are CC BY 4.0 (see LICENSE.md). Patient-level CSV files are not redistributed; data/README.md documents the source, schema, cleaning rules, and checksums. This is a retrospective research benchmark and not a clinical device.
Institution: Khulna University of Engineering and Technology
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7204268595