Getting Started
The sections below provide a high level overview of the CausalForge
package. This page takes you through installation, dependencies, main features, imputation methods supported, and basic usage of the package. It also provides links to get in touch with the authors, review our lisence, and review how to contribute.
Installation
Requires tensorflow and pytorch installed. For example, for a local Mac enviroment without GPUs, you can create one with
conda env create -f env_mac.yml
, and you activate it withconda activate causalforge
.Download
CausalForge
withpip install causalforge
.If
pip
cached an older version, trypip install --no-cache-dir --upgrade causalforge
.If you want to work with the development branch, use the script below:
Development
git clone -b dev --single-branch https://github.com/anthem-ai/causalforge
cd causalforge
python setup.py install
Versions and Dependencies
Python 3.8+
Dependencies:
numpy>=1.18.5
scipy>=1.4.1
pandas
cython
statsmodels
scikit-learn
matplotlib
pymc
seaborn
tqdm
tensorflow
keras
torch
Methods Supported
Name |
Paper [Link] |
Journal/Conference |
---|---|---|
BCAUSS |
Journal of Biomedical Informatics 2023 |
|
Dragonnet |
Claudia Shi et al, Adapting Neural Networks for the Estimation of Treatment Effects |
NeurIPS 2019 |
BCAUS |
BMC Medical Research Methodology 2021 |
|
GANITE |
ICLR 2018 |
License
Distributed under the MIT license. See LICENSE for more information.