hadisinaee / causal-ml-course

causal-ml-course

This repository contains code and data for experimentation with causal machine learning techniques. It includes synthetic and real-world datasets as well as code for running experiments.

Prerequisites

To use this repository, you will need:

• Python 3
• The virtual environment module for Python
• R libraries as required by the [Causal Discovery Toolbox][cdt]

Available Data

There are two tar.gz files in this repository:

1. results_20perc_n15.tar.gz: This file contains the results of a sparse experiment with 15 nodes and 5 trials.
2. results_70perc_n15.tar.gz: This file contains the results of a dense experiment with 15 nodes and 5 trials.

Installation and Usage

To set up the necessary packages and environment, follow these steps:

1. Create a virtual environment:

python3 -m venv venv
pip install -r requirements.txt

You can then use 15_Nodes.ipynb file to run all the experiments. This is a python notebook that includes all of our experiments. You can change the variables, such as number of node nnode, number of trials or many other variables.

Graph Partitioning Data

The file partitioning_data.py contains code for running experiments on graph partitioning data. This file loads the required data from the fennel_output.csv file. To run this experiment, simply run: python3 ./partitioning_data.py