danieletramontano/Causal-Effect-Identification-in-LiNGAM-Models-with-Latent-Confounders

Code and Data for '' Causal Effect Identification in LiNGAM Models with Latent Confounders''.

Dependencies

Recreate environment:

conda env create -f requirements.yml
conda activate experiments
conda install -c anaconda ipykernel
python -m ipykernel install --user --name=experiments

Usage

Identifiability Certification

• The code to run the algorithms described in the paper can be found in Certification/cert_utils.ipynb.
• To reproduce the figures in the paper, please see the Jupiter Notebook Certification/Identifiability_plots.ipynb.

Causal Effect Estimation

• The algorithms for estimation of the causal effects implemented in Estimation/src/python and Estimation/src/matlab.
• To reproduce the results one can run the jupyter-notebooks Estimation/main_laplace.ipynb and Estimation/main_exponential.ipynb. Please note that:
  • RICA algorithm is implemented in matlab. Therefore installed matlab is required to reproduce all the results. Please change the value of variable RUN_MATLAB to False in jupyter-notebooks to reproduce the results of all other algorithms except RICA.

License

See LICENSE for more information.

Authors

Daniele Tramontano, Yaroslav Kivva, Saber Salehkaleybar, Mathias Drton, Negar Kiyavash.