Research map / Probabilistic and scientific modeling / Treatment effects
A Tree-based Model Averaging Approach for Personalized Treatment Effect Estimation from Heterogeneous Data Sources
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| adaptively weight local CATE models using trees and site/subject features | CATE estimation with EF · Simulation, discrete grouping, c=1, n=500 | MSE ratio over LOC: 0.13 | Helped |
| adaptively weight local CATE models using trees and site/subject features | CATE estimation with EF · Simulation, continuous grouping, c=1, n=500 | MSE ratio over LOC: 0.3 | Helped |
| use an ensemble forest rather than an ensemble tree (ablation) | Tree-based model averaging · Simulation with large global heterogeneity | MSE | Helped |
| use ground-truth local CATEs to estimate ensemble weights (ablation) | EF · Simulation, discrete grouping, c=1, n=500 | MSE ratio over LOC: 0.13 -> <0.01 | Helped |
| use ground-truth local CATEs to estimate ensemble weights (ablation) | EF · Simulation, continuous grouping, c=1, n=500 | MSE ratio over LOC: 0.3 -> 0.05 | Helped |
| use ground-truth local CATEs to estimate ensemble weights (ablation) | ET · Simulation, discrete grouping, c=1, n=500 | MSE ratio over LOC: 0.16 -> <0.01 | Helped |
| use ground-truth local CATEs to estimate ensemble weights (ablation) | ET · Simulation, continuous grouping, c=1, n=500 | MSE ratio over LOC: 0.37 -> 0.07 | Helped |
| increase sample size at each site (ablation) | EF · Simulation, discrete grouping | MSE ratio over LOC: 0.57 -> 0.12 -> 0.07 | Helped |
| increase covariate dimension (ablation) | Tree-based model averaging · Simulation, D=20 or D=50 | MSE ratio over LOC | Hurt |
| use unequal site sample sizes, with 200 observations at external sites instead of 500 (ablation) | Tree-based model averaging · Simulation, site 1 n=500 | MSE ratio over LOC | Hurt |
| use causal forest instead of causal tree for local CATE models (ablation) | Tree-based model averaging · Simulation | MSE | No clear effect |
| adjust augmented-data observation weights for site sample size (ablation) | EF in the hospital application · eICU oxygen-therapy survival | expected survival improvement over baseline: 3% -> 4% | Helped |
Closest papers by meaning
- Estimating Heterogeneous Treatment Effects by Combining Weak Instruments and Observational Data (NeurIPS 2024 · similarity 0.88)
- Meta-Learners for Partially-Identified Treatment Effects Across Multiple Environments (ICML 2024 · similarity 0.87)
- A Non-parametric Direct Learning Approach to Heterogeneous Treatment Effect Estimation under Unmeasured Confounding (NeurIPS 2024 · similarity 0.87)
- Heterogeneous Treatment Effect in Time-to-Event Outcomes: Harnessing Censored Data with Recursively Imputed Trees (ICML 2025 · similarity 0.86)
- Differentially private learners for heterogeneous treatment effects (ICLR 2025 · similarity 0.86)
- PairNet: Training with Observed Pairs to Estimate Individual Treatment Effect (ICML 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.