csquires / dct-policy

This repository contains the code for the paper "Active Causal Structure Learning via Directed Clique Trees"

The following command will set up a Python virtual environment with the necessary dependencies installed.

bash setup.sh

To run the code, activate the virtual environment:

source venv/bin/activate

Then,

python3 fig1a.py
python3 fig1b.py
python3 fig1c.py

This will create two new directories. data/ will contain synthetic DAGs and the results (number of interventions used and computation time) for each intervention policy. figures/ will contain plots of these results as they appear in the paper.