Quantization for nonparametric regression

L. Györfi, Marten Wegkamp

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

The authors discuss quantization or clustering of nonparametric regression estimates. The main tools developed are oracle inequalities for the rate of convergence of constrained least squares estimates. These inequalities yield fast rates for both nonparametric (unconstrained) least squares regression and clustering of partition regression estimates and plug-in empirical quantizers. The bounds on the rate of convergence generalize known results for bounded errors to subGaussian, too.

Original languageEnglish
Pages (from-to)867-874
Number of pages8
JournalIEEE Transactions on Information Theory
Volume54
Issue number2
DOIs
Publication statusPublished - Feb 2008

Fingerprint

regression

Keywords

  • Finite-sample bounds
  • Least squares estimates
  • Regression estimation with restriction
  • Vector quantization

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Information Systems

Cite this

Quantization for nonparametric regression. / Györfi, L.; Wegkamp, Marten.

In: IEEE Transactions on Information Theory, Vol. 54, No. 2, 02.2008, p. 867-874.

Research output: Contribution to journalArticle

Györfi, L. ; Wegkamp, Marten. / Quantization for nonparametric regression. In: IEEE Transactions on Information Theory. 2008 ; Vol. 54, No. 2. pp. 867-874.
@article{574cc2dd2c3e40649f94b4f3cfb6d06a,
title = "Quantization for nonparametric regression",
abstract = "The authors discuss quantization or clustering of nonparametric regression estimates. The main tools developed are oracle inequalities for the rate of convergence of constrained least squares estimates. These inequalities yield fast rates for both nonparametric (unconstrained) least squares regression and clustering of partition regression estimates and plug-in empirical quantizers. The bounds on the rate of convergence generalize known results for bounded errors to subGaussian, too.",
keywords = "Finite-sample bounds, Least squares estimates, Regression estimation with restriction, Vector quantization",
author = "L. Gy{\"o}rfi and Marten Wegkamp",
year = "2008",
month = "2",
doi = "10.1109/TIT.2007.913565",
language = "English",
volume = "54",
pages = "867--874",
journal = "IEEE Transactions on Information Theory",
issn = "0018-9448",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "2",

}

TY - JOUR

T1 - Quantization for nonparametric regression

AU - Györfi, L.

AU - Wegkamp, Marten

PY - 2008/2

Y1 - 2008/2

N2 - The authors discuss quantization or clustering of nonparametric regression estimates. The main tools developed are oracle inequalities for the rate of convergence of constrained least squares estimates. These inequalities yield fast rates for both nonparametric (unconstrained) least squares regression and clustering of partition regression estimates and plug-in empirical quantizers. The bounds on the rate of convergence generalize known results for bounded errors to subGaussian, too.

AB - The authors discuss quantization or clustering of nonparametric regression estimates. The main tools developed are oracle inequalities for the rate of convergence of constrained least squares estimates. These inequalities yield fast rates for both nonparametric (unconstrained) least squares regression and clustering of partition regression estimates and plug-in empirical quantizers. The bounds on the rate of convergence generalize known results for bounded errors to subGaussian, too.

KW - Finite-sample bounds

KW - Least squares estimates

KW - Regression estimation with restriction

KW - Vector quantization

UR - http://www.scopus.com/inward/record.url?scp=39849110135&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=39849110135&partnerID=8YFLogxK

U2 - 10.1109/TIT.2007.913565

DO - 10.1109/TIT.2007.913565

M3 - Article

AN - SCOPUS:39849110135

VL - 54

SP - 867

EP - 874

JO - IEEE Transactions on Information Theory

JF - IEEE Transactions on Information Theory

SN - 0018-9448

IS - 2

ER -