Research map / Learning theory and algorithms / Kernels and regression theory
Model-free Variable Selection in Reproducing Kernel Hilbert Space
Lei Yang, Shaogao Lv, Junhui Wang
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 model-free gradient-based MF instead of additive-model selection (Add) | variable selection · Simulated Example 2, interaction (n=100,p=10,eta=0) | correct-fitting count out of 50: 25 -> 43 | Improved |
| use MF instead of Add when the true model is additive | variable selection · Simulated Example 1, additive (n=100,p=20,eta=0) | correct-fitting count out of 50: 48 vs 35 | No clear effect |
| use MF instead of Mars | variable selection · Simulated Example 1, additive (n=200,p=50,eta=0) | correct-fitting count out of 50: 0 -> 39 | Improved |
| use MF instead of SGL | variable selection · Simulated Example 1, additive (n=100,p=10,eta=0) | correct-fitting count out of 50: 12 -> 50 | Improved |
| add correlation among variables (eta 0 -> 1) | MF · Simulated Example 1, additive (n=100,p=10) | correct-fitting count out of 50: 50 -> 33 | Hurt |
| use MF instead of Add under correlated high-dimensional interaction data | variable selection · Simulated Example 2, interaction (n=200,p=50,eta=1) | averaged number of selected variables: 16.200 -> 2.160 | Improved |
| use MF instead of Cosso under interaction data | variable selection · Simulated Example 2, interaction (n=100,p=10,eta=0) | correct-fitting count out of 50: 15 -> 43 | Improved |
| use MF instead of Mars | prediction · Boston housing | averaged prediction error: 1.956 -> 1.774 | Improved |
| use MF instead of Add | prediction · Boston housing | averaged prediction error: 1.780 -> 1.774 | Improved |
| use MF instead of SGL | prediction · Ozone concentration | averaged prediction error: 1.776 -> 1.768 | Improved |
| use MF instead of Add (48 variables) | variable selection and prediction · Digit recognition (3 vs 5) | averaged prediction error: 1.871 -> 1.857 | Improved |
Closest papers by meaning
- Variable Selection via Penalized Neural Network: a Drop-Out-One Loss Approach (ICML 2018 · similarity 0.86)
- kernelPSI: a Post-Selection Inference Framework for Nonlinear Variable Selection (ICML 2019 · similarity 0.86)
- Differentially Private High-dimensional Variable Selection via Integer Programming (NeurIPS 2025 · similarity 0.86)
- Towards a Unified Framework for Uncertainty-aware Nonlinear Variable Selection with Theoretical Guarantees (NeurIPS 2022 · similarity 0.86)
- Gradient-Free Kernel Stein Discrepancy (NeurIPS 2023 · similarity 0.85)
- Kernel-based guarantees for nonlinear parametric models in Bayesian optimization (NeurIPS 2026 · 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.
