Recursive Partitioning for Heterogeneous Causal Effects

Abstract

In this paper we study the problems of estimating heterogeneity in causal effects in experimental or observational studies and conducting inference about the magnitude of the differences in treatment effects across subsets of the population. In applications, our method provides a data-driven approach to determine which subpopulations have large or small treatment effects and to test hypotheses about the differences in these effects. For experiments, our method allows researchers to identify heterogeneity in treatment effects that was not specified in a pre-analysis plan, without concern about invalidating inference due to multiple testing. In most of the literature on supervised machine learning (e.g. regression trees, random forests, LASSO, etc.), the goal is to build a model of the relationship between a unit's attributes and an observed outcome. A prominent role in these methods is played by cross-validation which compares predictions to actual outcomes in test samples, in order to select the level of complexity of the model that provides the best predictive power. Our method is closely related, but it differs in that it is tailored for predicting causal effects of a treatment rather than a unit's outcome. The challenge is that the "ground truth" for a causal effect is not observed for any individual unit: we observe the unit with the treatment, or without the treatment, but not both at the same time. Thus, it is not obvious how to use cross-validation to determine whether a causal effect has been accurately predicted. We propose several novel cross-validation criteria for this problem and demonstrate through simulations the conditions under which they perform better than standard methods for the problem of causal effects. We then apply the method to a large-scale field experiment re-ranking results on a search engine.

References in corpus (1)

• Doubly Robust Policy Evaluation and Learning

Cited by in corpus (135)

• Meta-learners for Estimating Heterogeneous Treatment Effects using Machine Learning
• Causal machine learning for predicting treatment outcomes
• A Review of Generalizability and Transportability
• 70 years of machine learning in geoscience in review
• Estimating individual treatment effect: generalization bounds and algorithms
• Machine Learning Estimation of Heterogeneous Causal Effects: Empirical Monte Carlo Evidence
• Double Machine Learning based Program Evaluation under Unconfoundedness
• High-dimensional regression adjustments in randomized experiments
• Machine learning for causal inference: on the use of cross-fit estimators
• Quasi-Oracle Estimation of Heterogeneous Treatment Effects
• Bayesian Inference of Individualized Treatment Effects using Multi-task Gaussian Processes
• Causal Decision Trees
• Heterogeneous Employment Effects of Job Search Programmes: A Machine Learning Approach
• Response Transformation and Profit Decomposition for Revenue Uplift Modeling
• Estimating the Effects of Continuous-valued Interventions using Generative Adversarial Networks
• Deep Counterfactual Networks with Propensity-Dropout
• Learning Weighted Representations for Generalization Across Designs
• Percent Change Estimation in Large Scale Online Experiments
• Counterfactual Prediction with Deep Instrumental Variables Networks
• Generalization Bounds and Representation Learning for Estimation of Potential Outcomes and Causal Effects
• Causal Inference in Travel Demand Modeling (and the lack thereof)
• Model Distillation for Revenue Optimization: Interpretable Personalized Pricing
• Learning to Prescribe Interventions for Tuberculosis Patients Using Digital Adherence Data
• Modern approaches for evaluating treatment effect heterogeneity from clinical trials and observational data
• Selective inference for effect modification via the lasso
• Experimental Evaluation of Individualized Treatment Rules
• Machine Learning Methods Economists Should Know About
• Bayesian Nonparametric Causal Inference: Information Rates and Learning Algorithms
• A Survey on Causal Inference
• Semiparametric Bayesian causal inference
• Guidelines for estimating causal effects in pragmatic randomized trials
• Global Model Interpretation via Recursive Partitioning
• Causal Interaction Trees: Tree-Based Subgroup Identification for Observational Data
• False Discovery Rate Controlled Heterogeneous Treatment Effect Detection for Online Controlled Experiments
• Transfer Learning for Estimating Causal Effects using Neural Networks
• Estimating heterogeneous treatment effects with right-censored data via causal survival forests
• Learning Triggers for Heterogeneous Treatment Effects
• Causal Falling Rule Lists
• Causal Reasoning in Software Quality Assurance: A Systematic Review
• Treatment effect estimation with disentangled latent factors
• Subgroup detection in linear growth curve models with generalized linear mixed model (GLMM) trees
• Learning Interpretable Models with Causal Guarantees
• Estimating causal effects with optimization-based methods: A review and empirical comparison
• Invariant Representation Learning for Treatment Effect Estimation
• Tree-Values: selective inference for regression trees
• Analysing a built-in advantage in asymmetric darts contests using causal machine learning
• A Tree-based Model Averaging Approach for Personalized Treatment Effect Estimation from Heterogeneous Data Sources
• Robust Recursive Partitioning for Heterogeneous Treatment Effects with Uncertainty Quantification
• Causal Effect Inference for Structured Treatments
• Off-Policy Evaluation of Ranking Policies under Diverse User Behavior
• Feature Selection Methods for Uplift Modeling and Heterogeneous Treatment Effect
• Learning When-to-Treat Policies
• Machine-Learning Tests for Effects on Multiple Outcomes
• Causal Inference with Latent Variables: Recent Advances and Future Prospectives
• SurvITE: Learning Heterogeneous Treatment Effects from Time-to-Event Data
• What Makes Forest-Based Heterogeneous Treatment Effect Estimators Work?
• Minimax Optimal Nonparametric Estimation of Heterogeneous Treatment Effects
• Inferring Individual Direct Causal Effects Under Heterogeneous Peer Influence
• Learning Optimal Policies from Observational Data
• Personalized Treatment Selection using Causal Heterogeneity
• Learning Treatment Effects in Panels with General Intervention Patterns
• Improving External Communication of Automated Vehicles Using Bayesian Optimization
• Estimating Structural Target Functions using Machine Learning and Influence Functions
• A general framework for causal classification
• A Comparison of Resampling and Recursive Partitioning Methods in Random Forest for Estimating the Asymptotic Variance Using the Infinitesimal Jackknife
• Targeting for long-term outcomes
• Latent-state models for precision medicine
• Achieving Reliable Causal Inference with Data-Mined Variables: A Random Forest Approach to the Measurement Error Problem
• Heterogeneous Effects of Software Patches in a Multiplayer Online Battle Arena Game
• FairUDT: Fairness-aware Uplift Decision Trees
• Flexible machine learning estimation of conditional average treatment effects: a blessing and a curse
• Variable importance measures for heterogeneous treatment effects
• A Groupwise Approach for Inferring Heterogeneous Treatment Effects in Causal Inference
• Uplift Modeling for Multiple Treatments with Cost Optimization
• Nonparametric Heterogeneous Treatment Effect Estimation in Repeated Cross Sectional Designs
• Gaussian Process Mixtures for Estimating Heterogeneous Treatment Effects
• Affordable Uplift: Supervised Randomization in Controlled Experiments
• Practical Guidance on Modeling Choices for the Virtual Twins Method
• Ordered Correlation Forest
• A unified survey of treatment effect heterogeneity modeling and uplift modeling
• Stable discovery of interpretable subgroups via calibration in causal studies
• Assessment of Heterogeneous Treatment Effect Estimation Accuracy via Matching
• Estimating Heterogeneous Causal Effects of High-Dimensional Treatments: Application to Conjoint Analysis
• Predicting Individual Treatment Effects of Large-scale Team Competitions in a Ride-sharing Economy
• Partial counterfactual identification and uplift modeling: theoretical results and real-world assessment
• Interpretable Assessment of Fairness During Model Evaluation
• A cautionary tale on fitting decision trees to data from additive models: generalization lower bounds
• Interval censored recursive forests
• Learning Continuous Treatment Policy and Bipartite Embeddings for Matching with Heterogeneous Causal Effects
• Heterogeneous Treatment Effect Estimation through Deep Learning
• Heterogeneous Causal Learning for Effectiveness Optimization in User Marketing
• Using Small Domain Estimation to obtain better retrospective Age Period Cohort insights
• What can the millions of random treatments in nonexperimental data reveal about causes?
• Generalised linear models for prognosis and intervention: Theory, practice, and implications for machine learning
• Inherent Trade-offs in the Fair Allocation of Treatments
• Improved inference for doubly robust estimators of heterogeneous treatment effects
• Doubly Robust Direct Learning for Estimating Conditional Average Treatment Effect
• Identification of Subgroups With Similar Benefits in Off-Policy Policy Evaluation
• Towards unique and unbiased causal effect estimation from data with hidden variables
• UTBoost: Gradient Boosted Decision Trees for Uplift Modeling
• Detecting Heterogeneous Treatment Effect with Instrumental Variables
• Identify treatment effect patterns for personalised decisions
• Building Trees for Probabilistic Prediction via Scoring Rules
• Applications of 0-1 Neural Networks in Prescription and Prediction
• Weighted Tensor Completion for Time-Series Causal Inference
• Enhanced inference for distributions and quantiles of individual treatment effects in various experiments
• Optimizing Treatment Allocation in the Presence of Interference
• Markdowns in E-Commerce Fresh Retail: A Counterfactual Prediction and Multi-Period Optimization Approach
• Finding and assessing treatment effect sweet spots in clinical trial data
• AutoCP: Automated Pipelines for Accurate Prediction Intervals
• Balancing Method for High Dimensional Causal Inference
• Counterfactual Learning to Rank using Heterogeneous Treatment Effect Estimation
• Eliciting Private User Information for Residential Demand Response
• Learning causal effects from many randomized experiments using regularized instrumental variables
• A New Transformation Approach for Uplift Modeling with Binary Outcome
• Estimating Heterogeneous Causal Effects in the Presence of Irregular Assignment Mechanisms
• Predictive Directions for Individualized Treatment Selection in Clinical Trials
• Unbiased Estimation of the Value of an Optimized Policy
• Approaching an unknown communication system by latent space exploration and causal inference
• Machine Learning for Experimental Design: Methods for Improved Blocking
• Heterogeneous Treatment Effects in Social Networks
• Efficient Balanced Treatment Assignments for Experimentation
• Modelling hetegeneous treatment effects by quantitle local polynomial decision tree and forest
• Estimating Potential Outcome Distributions with Collaborating Causal Networks
• Causal Transfer Random Forest: Combining Logged Data and Randomized Experiments for Robust Prediction
• Local search for efficient causal effect estimation
• Learning and Testing Sub-groups with Heterogeneous Treatment Effects:A Sequence of Two Studies
• Estimation of Discrete Choice Models: A Machine Learning Approach
• Horseshoe Forests for High-Dimensional Causal Survival Analysis
• A Statistical Test for the Benefits of Personalizing Interventions
• Estimation and Validation of Ratio-based Conditional Average Treatment Effects Using Observational Data
• Robust and Heterogenous Odds Ratio: Estimating Price Sensitivity for Unbought Items
• The Bias-Variance Tradeoff of Doubly Robust Estimator with Targeted $L_1$ regularized Neural Networks Predictions
• Discovering Context Specific Causal Relationships
• Machine learning reveals how personalized climate communication can both succeed and backfire