AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

[Re] Hierarchical Shrinkage: Improving the Accuracy and Interpretability of Tree-Based Methods

Domen Mohorčič, David Ocepek

NeurIPS 2023 · Architectures and efficiency › RNNs and normalization · approach: Learned embeddings for retrieval

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 changedWhereMeasuredEffect
Apply hierarchical shrinkage post-hoc to CART treesCART · Classification datasets, 15 leavesrelative AUC: 6.2Improved
Apply HS on top of cost-complexity pruningCART with CCP · Classification datasets, 15 leavesrelative AUC: 6.5Improved
Apply HS to C4.5 treesC4.5 · Classification datasets, 15 leavesrelative AUC: 8Improved
Apply HS to CART on regression dataCART · Regression datasets, 15 leavesrelative R2: 9.8Improved
Apply HS to CART with CCP on regression dataCART with CCP · Regression datasets, 15 leavesrelative R2: 10.1Improved
Replace leaf-based shrinkage with hierarchical shrinkage (ablation)CART · Classification and regression datasetsAUC and R2Improved
Increase number of leaves (deeper trees) with HS (ablation)CART and CART with CCP · Larger datasetsprediction performanceImproved
Apply HS to each tree of a Random ForestRandom Forest · Classification and regression datasetsprediction accuracyImproved
Regularize RF with HS instead of max depth or mtry (ablation)Random Forest · Classification datasetsprediction accuracyMixed
Apply HS to a two-feature RFRandom Forest · Diabetes classificationAUC: 0.733 -> 0.787Improved
Apply HS to BART treesBART · Classification and regression datasetsAUC and R2No clear effect
Apply HS to GOSDT optimal treesGOSDT · Selected classification datasetsAUCNo clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy