Research map / Learning theory and algorithms / Prediction and decision losses
Random Forests, Decision Trees, and Categorical Predictors: The "Absent Levels" Problem
Timothy C. Au
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 a missing data heuristic (best of Stop/Majority/Random/DBI) instead of always sending absent levels left | randomForest R package · 1985 Auto Imports | RMSE: 6.2% worse | Improved |
| use a missing data heuristic instead of One-Hot encoding | random forest · 1985 Auto Imports | RMSE: 2.2% worse | Improved |
| use a missing data heuristic instead of the Left heuristic for absent levels | randomForest R package · PROMESA | ROC AUC: 1.5% worse | Improved |
| use a missing data heuristic instead of the Left heuristic for absent levels | randomForest R package · PROMESA | PR AUC: 3.7% worse | Improved |
| use DBI heuristic rather than Stop, Majority or Random | random forest missing data heuristics · 1985 Auto Imports | RMSE | Improved |
| use Random heuristic rather than Majority, Stop or DBI | random forest missing data heuristics · PROMESA | PR AUC | Improved |
| use Majority heuristic rather than other missing data heuristics | random forest missing data heuristics · PROMESA | ROC AUC | Improved |
| use One-Hot encoding rather than missing data heuristics | random forest · PROMESA | PR AUC | Mixed |
| use Right heuristic (always send absent levels right) | random forest · 1985 Auto Imports | RMSE | No clear effect |
Closest papers by meaning
- Hierarchical Shrinkage: Improving the accuracy and interpretability of tree-based models. (ICML 2022 · similarity 0.90)
- Multistage Defer Trees for Hybrid Interpretability: If at First You Can't Succeed, Tree Again (NeurIPS 2026 · similarity 0.86)
- DynFrs: An Efficient Framework for Machine Unlearning in Random Forest (ICLR 2025 · similarity 0.86)
- Generalized Random Forests Using Fixed-Point Trees (ICML 2025 · similarity 0.85)
- Learning Small Decision Trees with Few Outliers: A Parameterized Perspective (AAAI 2024 · similarity 0.85)
- Symbolic Regression Enhanced Decision Trees for Classification Tasks (AAAI 2024 · similarity 0.85)
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.
