Less Greedy Equivalence Search is an algorithm for causal discovery from observational and interventional data with prior assumptions.

This repository contains the code for the paper "Less Greedy Equivalence Search" by Adiba Ejaz and Elias Bareinboim.

Disclaimer: This code is offered publicly with the MIT License, meaning that others can use it for any purpose, but we are not liable for any issues arising from the use of this code. Please cite our work if you found this code useful.

LGES/
├── ges 
├── src/
│   └── experiments/
│       ├── obs/        
│       │   ├── obs_no_prior.py
│       │   ├── obs_init.py
│       │   └── obs_example.py
│       └── exp/       
│           └── exp_no_prior.py
│       └── sachs.py
└── README.md

Make sure you have Python 3.8+ and the required dependencies installed.

pip install -r requirements.txt

cd src/experiments/obs
python -m obs_no_prior \
    -p <NUM_VARS> \
    -n <NUM_SAMPLES> \
    -d <AVG_DEGREE> \
    -g <NUM_TRIALS> \
    -m <MODEL_TYPE>

Results saved to:

src/experiments/obs/results_no_prior/vars_<NUM_VARS>_samples_<NUM_SAMPLES>_degree_<AVG_DEGREE>_model_<MODEL_TYPE>

cd src/experiments/obs
python -m obs_init

Results saved to:

src/experiments/obs/results_init/vars_<NUM_VARS>_samples_<NUM_SAMPLES>_degree_<AVG_DEGREE>_model_<MODEL_TYPE>

cd src/experiments/exp
python -m exp_no_prior \
    -p <NUM_VARS> \
    -n <NUM_SAMPLES> \
    -d <AVG_DEGREE> \
    -g <NUM_TRIALS> \
    -m <MODEL_TYPE>

Results saved to:

src/experiments/exp/results_no_prior/vars_<NUM_VARS>_samples_<NUM_SAMPLES>_degree_<AVG_DEGREE>_model_<MODEL_TYPE>

To generate the plots presented in the paper, use:

cd src/experiments/obs
python -m plot.py --plot obs_no_prior
python -m plot.py --plot obs_init
cd src/experiments/exp
python -m plot.py --plot exp_no_prior

Plots saved to:

src/experiments/obs/plots_no_prior/
src/experiments/obs/plots_init/
src/experiments/exp/plots_no_prior/

To see a minimal working example:

# src/experiments/obs/obs_example.py

This shows how to configure and run a basic experiment.