WBIC model selection in Bayesian Mixed LiNGAM
This program is under development.

Python Library for statistical causal inference based on the Bayesian Mixed LiNGAM (linear non-gaussian acyclic model) and WBIC (Widely applicable Bayesian Information Criterion).

Please install the developed version of bmlingam before installation of wbic_bml.

git clone -b develop https://github.com/taku-y/bmlingam
cd bmlingam
python setup.py install

Then install wbic_bml.

git clone https://github.com/inoueakimitsu/wbic_bml.git
cd wbic_bml
python setup.py install

Please refer to the test code in tests directory.

wbic_bml is tested on Python 3.5 and depends on PyMC3 3.2.

• Shimizu, S., & Bollen, K. (2014). Bayesian estimation of causal direction in acyclic structural equation models with individual-specific confounder variables and non-Gaussian distributions. Journal of Machine Learning Research, 15(1), 2629-2652.