Example: Scikit-learn baseline

This example fits a logistic regression classifier.

What this is / is not - Is: A very fast baseline with minimal dependencies. - Is not: DP or FL. Use this for sanity checks and baselines. - Is not: Orchestration. For Aegis API flows, see technical/train_model.md.

When to use - Sanity-check modeling pipeline or compare to DP/FL runs.

Prerequisites - Python deps:

pip install scikit-learn

Run it

python examples/sklearn_demo.py --n 500 --d 5 --test_size 0.2 --seed 0

Output (sample)

Train acc: 0.92 | Test acc: 0.90

Notes - Great for a quick sanity baseline before DP/FL.