Federated Computation of ROC and PR Curves

Installation

1. Create a Python environment:

conda create --name fedcurve python=3.11
conda activate fedcurve

2. Clone the repository:

git clone https://github.com/xuefeng-xu/fedcurve.git && cd fedcurve

3. Install dependencies:

pip install -e .

Usage

See fedcurve/README for details.

Reproduction

To reproduce results for the adult dataset with XGBClassifier:

python reproduce.py --dataset adult --classifier XGBClassifier

Similarly, to reproduces results for the bank dataset with LogisticRegression:

python reproduce.py --dataset bank --classifier LogisticRegression

Plots are saved under ./img/*/{dataset}_{classifier}_{curve}.pdf