mrsamsami/Causality

Causal analysis and inference using observational and interventional dataset. It contains tools for graph structure recovery.

Causality

This repository provides implemented algorithms for causal inference between two variables and learning causal structures in graphs. All code is written in Python 3, using SciPy, NumPy, Scikit-learn, Hyppo, and NetworkX. Jupyter notebooks are provided to show analyses.

Note: These are implemented algorithms and study for assignments of Causal Inference taught by Prof. Salehkalyebar in Fall 2020. This repository is in its early phases.

Table of Contents

• Causal Inference Between Two Variables
• Learning Causal Skeleton in Multivariate Setting
• Calculating Size of the Markov Equivalence Class

Environment Requirement

• Python 3.6.5
  • NumPy 1.14.3
  • SciPy 1.5.4
  • Scikit-learn 0.23.2
  • Hyppo 0.1.3
  • NetworkX 2.5

Resources

Class:

• Causal Inference (Sharif, Fall 2020)

Textbooks:

• Elements of Causal Inference: Foundations and Learning Algorithms (2018)
• Causality: Models, Reasoning and Inference (2009)

Paper:

• A linear non-gaussian acyclic model for causal discovery (2006)