Research map / Learning theory and algorithms / Prediction and decision losses
Tunability: Importance of Hyperparameters of Machine Learning Algorithms
Philipp Probst, Anne-Laure Boulesteix, Bernd Bischl
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 |
|---|---|---|---|
| replace package defaults with data-derived optimal defaults | glmnet · OpenML binary (AUC) | tunability: 0.069 -> 0.024 | Improved |
| replace package defaults with optimal defaults | rpart · OpenML binary (AUC) | tunability: 0.038 -> 0.012 | Improved |
| replace package defaults with optimal defaults | kknn · OpenML binary (AUC) | tunability: 0.031 -> 0.006 | Improved |
| replace package defaults with optimal defaults | svm · OpenML binary (AUC) | tunability: 0.056 -> 0.042 | Improved |
| replace package defaults with optimal defaults | ranger · OpenML binary (AUC) | tunability: 0.010 -> 0.006 | Improved |
| replace package defaults with optimal defaults | xgboost · OpenML binary (AUC) | tunability: 0.043 -> 0.014 | Improved |
| cross-validate optimal defaults | svm · OpenML binary (accuracy) | cross-validated improvement: 0.011 -> -0.001 | No clear effect |
| tune two hyperparameters jointly (minbucket and minsplit) instead of the best one | rpart · OpenML binary (AUC) | joint gain: 0.0055 | Improved |
| tune cp jointly with maxdepth instead of cp alone | rpart · OpenML binary (AUC) | joint gain: 0.0007 | No clear effect |
| tune xgboost eta or booster | xgboost · OpenML binary (AUC) | tunability: 0.005 and 0.008 | Improved |
Closest papers by meaning
- Weighted Sampling for Combined Model Selection and Hyperparameter Tuning (AAAI 2020 · similarity 0.84)
- Sample complexity of data-driven tuning of model hyperparameters in neural networks with structured parameter-dependent dual function (NeurIPS 2025 · similarity 0.84)
- MetaOptimize: A Framework for Optimizing Step Sizes and Other Meta-parameters (ICML 2025 · similarity 0.82)
- Taming Hyperparameter Sensitivity in Data Attribution: Practical Selection Without Costly Retraining (NeurIPS 2025 · similarity 0.82)
- A Survey of Learning Criteria Going beyond the Usual Risk (Abstract Reprint) (AAAI 2024 · similarity 0.81)
- Reproduce, Replicate, Reevaluate. The Long but Safe Way to Extend Machine Learning Methods (AAAI 2024 · similarity 0.81)
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.
