Supplementary Materials
Manuscript: Revisiting Differentially Private Federated Learning for Tabular Data:
A Matched-Accounting Benchmark of Boosting versus DP-SGD (electronics-4432478)

Contents
--------
scripts/   Self-contained experiment scripts (Python 3.10+, NumPy/SciPy/dp-accounting):
           tstar_simulation.py      Appendix A.4 (Figure A1, Table A1); run stages 1, 2, 3
           mia_experiment.py        Section 5.5 (Figure 4, Table 9)
           heterogeneity_sweep.py   Section 5.6 (Figure 5, Table 10)
           profiling_and_rf.py      Section 6.2 (Tables 11-12) and Appendix B (Table A2)
           dp_learners.py           Shared learner implementations (Sections 3.4-3.5 semantics)

results/   Result logs quoted in the manuscript (CSV/JSON), including the
           PLD-versus-RDP calibration verification data (pld_verification.json).

requirements.txt lists the minimum software versions.

Setup:  pip install -r requirements.txt
All scripts are deterministic (fixed seeds) and run on a single CPU core;
outputs are written to the working directory. The complete repository is at
https://github.com/asalzahrani-lab/dp-fl-tabular-benchmark
(archived: https://doi.org/10.5281/zenodo.21711522).
