AI Paper Atlas

Research map / Learning theory and algorithms / Kernels and regression theory

Dimension Reduction and MARS

Yu Liu LIU, Degui Li, Yingcun Xia

JMLR 2023 · Learning theory and algorithms › Kernels and regression theory · approach: High-dimensional linear regression theory

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
replace MARS on raw covariates with SDR directions estimated from MARS gradients (drMARS)MARS · data.3 Parkinsons, train 1000rMSPE (%): 32.16 -> 12.79Improved
replace MARS on raw covariates with SDR directions estimated from MARS gradients (drMARS)MARS · data.6 Hill-Valley, train 404MCR (%): 19.38 -> 6.21Improved
replace MARS on raw covariates with SDR directions (drMARS)MARS · Simulation M1, p=50, n=500, uniform XMSE(G): 0.52 -> 0.09Improved
use drMARS instead of MARS on data.2MARS · data.2 house prices, train 1000rMSPE (%): 15.50 -> 14.62Improved
use drMARS instead of MARS on pistachio classificationMARS · data.5 pistachio, train 716MCR (%): 9.30 -> 9.07No clear effect
use random directions instead of estimated SDR directions (ablation)SDR-MARS · data.1 concrete, train 343rMSPE (%): 15.85 -> 21.42Hurt
estimate SDR directions with pHd instead of drMARS (ablation)SDR-MARS · data.4 residential building, train 124rMSPE (%): 5.85 -> 12.82Hurt
estimate SDR directions with drMARS instead of MAVESDR estimation · Simulation M4, X~N, p=50, n=500SDR error D: 0.82 -> 0.04Improved
use drMARS instead of MAVE for SDR estimation under uniform XSDR estimation · Simulation M5, X~U, p=50, n=500SDR error D: 0.11 -> 0.49Hurt
use drMARS instead of CVE for SDR computing timeSDR estimation · Simulation M1, X~N, p=100, n=500computing time (s): 55.85 -> 6.43Improved
use drMARS instead of MAVE for choosing the true SDR dimensionSDR dimension selection · Simulation M7, X~N, p=100, n=500proportion selecting true dimension: 0.07 -> 0.53Improved
use drMARS instead of RF as regression estimatorRF · Simulation M1, p=50, n=500, uniform XMSE(G): 0.61 -> 0.09Improved

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