Reprocuible Package for Interpretable ML in Cardiovascular Risk Prediction (v1.0.0)

This deposit contains a fully reproducible package for our cardiovascular risk-prediction study. It includes code (src/, scripts/), configurations (configs/), processed data (data/processed/, no raw data), and evaluation outputs (outputs/tables/, outputs/figures/, outputs/runinfo/, outputs/splits/). Data availability. Raw data are not redistributed due to licensing. Retrieve the upstream dataset from Kaggle (Cardiovascular Diseases Risk Prediction Dataset, Alphiree; upstream source: CDC BRFSS 2021). We provide scripts/fetch_kaggle_cvd.sh and docs/data_source_kaggle.md with exact instructions. Reproduction. Create a Python environment, install requirements.txt, retrieve raw data with the provided script, run preprocessing (src/05_preprocess.py), splitting, model training, and evaluation (discrimination, calibration & Murphy decomposition, decision curve analysis, subgroup audits). All commands and seeds are documented in runbook.md and outputs/runinfo/. Contents. MANIFEST.json, CHECKSUMS.txt, code, configs, processed data, figures/tables used in the manuscript, and optional model artifacts (if included). Licensing. See the deposit license below and included LICENSE. Code is open-source; documentation is CC-BY; raw data are excluded.