Research map / Learning theory and algorithms / Prediction and decision losses
Tree-Based Models for Correlated Data
Assaf Rabinowicz, Saharon Rosset
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Use RETCO (correlation-aware splitting with bias-corrected error) instead of standard regression tree | Regression tree · Simulation, new random effects | error difference vs standard (%): -2.5, -9.9 and -25.2 for sigma_b^2 0.5, 1, 5 | Improved |
| Use RETCO instead of standard regression tree when random effects are shared with training | Regression tree · Simulation, same random effects | error difference vs standard (%): -2.1, -1.1 and -10.4 | Improved |
| Enforce the stopping rule constraint in RETCO random forest (ablation) | RETCO random forest · Simulation RF, new random effects | error difference vs standard (%): -2.1, -2.9, -11.7 -> -2.2, -9.6, -34.7 | Improved |
| Increase cluster size n_c from 50 to 150 | RETCO regression tree · Simulation, in-sample, new random effects | error difference vs standard (%): -9.9 -> -4.3 | Hurt |
| Use RETCO instead of standard tree under generalization error (CV_c vs CV) | Regression tree · Simulation, generalization error | error difference vs standard (%): -2.8, -11.0 and -7.3 | Improved |
| Use RETCO instead of RE-EM | RETCO · Simulation vs RE-EM | error difference vs standard (%) | Improved |
| Replace training error by bias-corrected prediction error in splitting (RETCO) (ablation) | Regression tree · Simulation, homogeneity analysis | error difference vs standard (%) | Improved |
| Use RETCO instead of standard regression tree on real data | Regression tree · Six real datasets | error difference vs standard (%): -4.9% to -32.9% | Improved |
| Use RETCO random forest instead of standard random forest on real data | Random forest · Six real datasets | error difference vs standard (%): -3.3% to -35.4% | Improved |
| Use RETCO iterative fitting instead of standard recursive tree fitting | RETCO · Simulation run time | run time: 3-8 fold longer | Hurt |
Closest papers by meaning
- Regression Trees Know Calculus (NeurIPS 2025 · similarity 0.87)
- Multivariate Boosted Trees and Applications to Forecasting and Control (JMLR 2022 · similarity 0.87)
- Smooth And Consistent Probabilistic Regression Trees (NeurIPS 2020 · similarity 0.86)
- Extrapolated Random Tree for Regression (ICML 2023 · similarity 0.86)
- Piecewise Constant and Linear Regression Trees: An Optimal Dynamic Programming Approach (ICML 2024 · similarity 0.85)
- TreeVI: Reparameterizable Tree-structured Variational Inference for Instance-level Correlation Capturing (NeurIPS 2024 · similarity 0.84)
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.
