Research map / Learning theory and algorithms / Prediction and decision losses
Optimal Weighted Random Forests
Xinyu Chen, Dalei Yu, Xinyu Zhang
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 optimal Mallows-type weights instead of equal weights | random forest with CART trees · UCI YH (CART trees) | MSFE: 35.442 -> 3.727 | Improved |
| use optimal Mallows-type weights instead of equal weights | random forest with SUT trees · UCI YH (SUT trees) | MSFE: 33.241 -> 2.433 | Improved |
| use optimal Mallows-type weights instead of equal weights | random forest with SUT trees · UCI BH (SUT trees) | MSFE: 38.213 -> 24.516 | Improved |
| use optimal weights instead of equal weights | random forest with CART trees · 12 UCI data sets (CART trees) | MSFE: essential improvement in 10 out of 12 data sets | Improved |
| use optimal weights instead of equal weights | random forest with SUT trees · 12 UCI data sets (SUT trees) | MSFE: 11 out of 12 data sets | Improved |
| use two-step quadratic optimization instead of one-step nonlinear optimization for weights | WRF_opt with CART trees · UCI PT | running time (s): 40.445 -> 0.291 | Improved |
| use two-step quadratic optimization instead of one-step nonlinear optimization for weights | WRF_opt with CART trees · UCI CCPP | running time (s): 2.785 -> 0.566 | Improved |
| use two-step instead of one-step weight optimization | WRF_opt with CART trees · UCI BH (CART trees) | MSFE: 14.038 -> 13.958 | No clear effect |
| use WRF_opt weights instead of wRF weights | random forest with CART trees · UCI AM (CART trees) | MSFE: 9.077 vs 9.272 | Hurt |
| use WRF_opt weights instead of wRF weights | random forest with CART trees · UCI QSAR (CART trees) | MSFE: 1.417 vs 1.423 | Hurt |
| use WRF_opt weights instead of CRF weights | random forest with CART trees · UCI YH (CART trees) | MSFE: 13.422 -> 3.727 | Improved |
| replace CART trees by SUT trees (unsupervised splits) for WRF_opt | WRF_opt · UCI BH | MSFE: 13.958 -> 24.516 | Hurt |
Closest papers by meaning
- Randomization as Regularization: A Degrees of Freedom Explanation for Random Forest Success (JMLR 2020 · similarity 0.87)
- Distributional Random Forests: Heterogeneity Adjustment and Multivariate Distributional Regression (JMLR 2022 · similarity 0.86)
- Hierarchical Shrinkage: Improving the accuracy and interpretability of tree-based models. (ICML 2022 · similarity 0.86)
- Comparison-Based Random Forests (ICML 2018 · similarity 0.86)
- Randomization Can Reduce Both Bias and Variance: A Case Study in Random Forests (JMLR 2025 · similarity 0.86)
- Generalized Random Forests Using Fixed-Point Trees (ICML 2025 · similarity 0.83)
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.
