/PGM-Causal-Reasoning

Experiments on structure learning of Bayesian Networks with emphasis on finding causal relationship

Primary LanguageJupyter NotebookMIT LicenseMIT

PGM-Causal-Reasoning

Experiments on structure learning of Bayesian Networks with emphasis on finding causal relationship

These experiments are done in context of the course Probabilistic Graphical Models (WiSe 18/19) held by Prof. Kristian Kersting at at the Technische Universität Darmstadt.

Datasets

Artifical generated data

LUCAS (LUng CAncer Simple set)

Real world data

CINA-dataset (Census Is Not Adult dataset)

Frameworks

SPFlow - An Easy and Extensible Library for Sum-Product Networks

bnlearn - An R package for Bayesian network learning and inference

BDgraph: Bayesian Structure Learning in Graphical Models using Birth-Death MCMC (R)

BNFinder - Tool for learning bayesian networks

Note: I needed to install a slightly older version for scipy in order to avoid: scipy/scipy#9606

SyntaxError: Non-ASCII character '\xe2'

using

sudo pip2 install scipy==1.1.0

libpgm - An endeavor to make Bayesian probability graphs easy to use

BayesSpy – Bayesian Python

OpenGM - a C++ template library for discrete factor graph models and distributive operations

Citation

If you find this project useful please consider citing:

@misc{queensgambit_experiments_2019,
	title = {Experiments on structure learning of {Bayesian} {Networks} with emphasis on finding causal relationship: {QueensGambit}/{PGM}-{Causal}-{Reasoning}},
	copyright = {MIT},
	shorttitle = {Experiments on structure learning of {Bayesian} {Networks} with emphasis on finding causal relationship},
	url = {https://github.com/QueensGambit/PGM-Causal-Reasoning},
	urldate = {2019-03-15},
	author = {Willig, Moritz and Czech, Johannes},
	month = feb,
	year = {2019},
	note = {original-date: 2019-01-20T13:56:16Z}
}

Our project report can be found here.

References

Graphical Models for Probabilistic and Causal Reasoning, Judea Pearl

Towards A Rigorous Science of Interpretable Machine Learning, Finale Doshi-Velez, Been Kim

Partial orientation and local structural learning of causal networks for prediction, Jianxin Yin, You Zhou, Changzhang Wang, Ping He, Cheng Zheng, Zhi Geng

An Exploration of Structure Learning in Bayesian Networks, Constantin Berzan

Learning Bayesian Networks with the bnlearn R Package