EconML

EconML is a Python package that uses machine learning to estimate heterogeneous treatment effects (HTE) and individualized causal responses.

EconML, developed by the ALICE team at Microsoft Research, is your go-to Python library for advanced causal inference. It merges econometric rigor with machine learning flexibility to estimate Conditional Average Treatment Effects (CATE), moving past single-number Average Treatment Effects (ATE). The package provides a unified API for state-of-the-art methods, including Double Machine Learning (DML), Orthogonal Random Forest (ORF), and meta-learners (S-, T-, X-Learners). This toolkit delivers reliable, interpretable 'what-if' predictions, enabling personalized decisions: for example, identifying which customer segments will respond most to a specific price discount.

https://pywhy.org/econml/