Research map / Learning theory and algorithms / Sample complexity
Non-parametric Quantile Regression via the K-NN Fused Lasso
Steven Siwei Ye, Oscar Hernan Madrid Padilla
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 quantile K-NN fused lasso instead of quantile random forest under Cauchy errors | Scenario 1, Cauchy errors, n=10000 · Simulation | MSE: 57748.95 -> 0.0962 | Improved |
| use quantile K-NN fused lasso instead of K-NN fused lasso under Cauchy errors | Scenario 1, Cauchy errors, n=10000 · Simulation | MSE: 4799.0 -> 0.0962 | Improved |
| use quantile K-NN fused lasso instead of K-NN fused lasso under Gaussian errors | Scenario 1, N(0,1), n=10000 · Simulation | MSE: 0.0292 -> 0.0370 | Hurt |
| use quantile K-NN fused lasso instead of quantile random forest on Scenario 3 (t2 errors) | Scenario 3, n=10000 · Simulation | MSE: 1.2910 -> 0.0059 | Improved |
| use quantile K-NN fused lasso instead of K-NN fused lasso on Scenario 2 (t3 errors) | Scenario 2, n=10000 · Simulation | MSE: 0.0542 -> 0.0265 | Improved |
| use quantile K-NN fused lasso instead of quantile random forest at small sample size, Scenario 4 tau=0.9 | Scenario 4, n=100 · Simulation | MSE: 0.6948 -> 0.7413 | Hurt |
| use quantile K-NN fused lasso instead of quantile random forest at larger sample size, Scenario 4 tau=0.9 | Scenario 4, n=1000 · Simulation | MSE: 0.5374 -> 0.3568 | Improved |
| use quantile K-NN fused lasso instead of quantile random forest for median | California housing, train size 10000 · California housing | MSE: 0.1623 -> 0.1479 | Improved |
| use quantile K-NN fused lasso instead of quantile random forest for 95% interval coverage | California housing, train size 10000 · California housing | coverage proportion: 0.8902 -> 0.9274 | Improved |
| use quantile K-NN fused lasso instead of quantile random forest for median, small training size | Chicago crime, train size 500 · Chicago crime | MSE: 1.1441 -> 1.2161 | Hurt |
| use quantile K-NN fused lasso instead of quantile random forest for median, large training size | Chicago crime, train size 2000 · Chicago crime | MSE: 0.8383 -> 0.8476 | No clear effect |
| use majorize-minimize instead of linear programming or ADMM solver | quantile K-NN fused lasso · Scenario 3 timing | computation time | Improved |
Closest papers by meaning
- Online Quantile Regression (JMLR 2025 · similarity 0.90)
- Additive Nonlinear Quantile Regression in Ultra-high Dimension (JMLR 2022 · similarity 0.89)
- Communication-Constrained Distributed Quantile Regression with Optimal Statistical Guarantees (JMLR 2022 · similarity 0.89)
- Confidence Intervals and Hypothesis Testing for High-dimensional Quantile Regression: Convolution Smoothing and Debiasing (JMLR 2023 · similarity 0.89)
- Distributed High-Dimensional Quantile Regression: Estimation Efficiency and Support Recovery (ICML 2024 · similarity 0.89)
- Relaxed Quantile Regression: Prediction Intervals for Asymmetric Noise (ICML 2024 · similarity 0.86)
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.
