Estimation Methods under Unconfoundedness

This section contains methods for estimating (heterogeneous) treatment effects, whose theoretical guarantees are valid only when all potential confounders/controls (factors that simultaneously had a direct effect on the treatment decision in the collected data and the observed outcome) are observed.

• Orthogonal/Double Machine Learning
  • What is it?
  • What are the relevant estimator classes?
  • When should you use it?
  • Overview of Formal Methodology
  • Class Hierarchy Structure
  • Usage FAQs
  • Usage Examples
• Doubly Robust Learning
  • What is it?
  • What are the relevant estimator classes?
  • When should you use it?
  • Overview of Formal Methodology
  • Class Hierarchy Structure
  • Usage FAQs
  • Usage Examples
• Forest Based Estimators
  • What is it?
  • What are the relevant estimator classes?
  • When should you use it?
  • Overview of Formal Methodology
    • Orthogonal Random Forests
    • CausalForest (aka Forest Double Machine Learning)
    • Forest Doubly Robust Learner
  • Class Hierarchy Structure
  • Usage Examples
• Meta-Learners
  • What is it?
  • What are the relevant estimator classes?
  • When should you use it?
  • Overview of Formal Methodology
    • T-Learner
    • S-Learner
    • X-Learner
    • Domain Adaptation Learner
    • Doubly Robust Learner
    • Non-Parametric Double Machine Learning
  • Class Hierarchy Structure
  • Usage Examples