Locally differentially Private Decision Tree (LPDT)

This repository implements LPDT, the Locally differentially Private Decision Tree described in the paper Decision Tree for Locally Private Estimation with Public Data accepted for NeurIPS 2023. The implementation is based on pure Python with the following required packages:

• Scikit-learn
• NumPy
• Numba
• Scipy

Contents

• Installation
• Demo
• References

Installation

Via PyPI

pip install LPDT

Via GitHub

pip install git+https://github.com/Karlmyh/LPDT.git

Manual Install

git clone git@github.com:Karlmyh/LPDT.git
cd LPDT 
python setup.py install

Demo

See simulation.py for a demo of the use of the class.

References

• Decision Tree for Locally Private Estimation with Public Data (NeurIPS 2023)