CausalML: Python Package for Causal Machine Learning

TL;DR

The paper addresses the need for practical, scalable causal inference tools that operate at the individual level. It introduces CausalML, a Python package offering eight uplift models and utilities to estimate $CATE$ and uplift curves, with support for multiple outcome types and multi-arm experiments. The work contributes a one-stop, open-source toolkit that democratizes uplift modeling and provides in-house meta-learners for optimizing multiple treatments. This enables targeting optimization, causal impact analysis, and personalization in real-world settings, while noting that randomized experiments remain essential for reliable ATE estimation.

Abstract

CausalML is a Python implementation of algorithms related to causal inference and machine learning. Algorithms combining causal inference and machine learning have been a trending topic in recent years. This package tries to bridge the gap between theoretical work on methodology and practical applications by making a collection of methods in this field available in Python. This paper introduces the key concepts, scope, and use cases of this package.

Figures (1)

• Figure 1: CausalML package diagram