Explaining Heterogeneous Treatment Effects with Causal Forests and the Hoeffding Functional Decomposition
To address this limitation, we introduce CausalTreeHFD, an interpretable extension of causal forests, based on the Hoeffding functional decomposition adapted to tree ensembles.
ProofPaper ↗
Key points
- The estimation of heterogeneous treatment effects is a central objective of causal inference.
- Consequently, the identification of covariates driving heterogeneity remains a challenging and open problem.
- Our approach decomposes predictions of conditional average treatment effects (CATE) into contributions from individual covariates and pairwise interactions, providing a direct characterization of the factors driving treatment effect heterogeneity.
- Experiments on simulated and real datasets show that CausalTreeHFD accurately identifies relevant covariates and interactions, and precisely recovers their contributions to heterogeneous treatment effects.
Sources (1)
- [1]Explaining Heterogeneous Treatment Effects with Causal Forests and the Hoeffding Functional DecompositionarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 12, 09:14 AM
To address this limitation, we introduce CausalTreeHFD, an interpretable extension of causal forests, based on the Hoeffding functional decomposition adapted to tree ensembles.
The estimation of heterogeneous treatment effects is a central objective of causal inference.
Extractive summary: sentences quoted from the sources.