Quantization for nonparametric regression

被引：6

作者：

Gyoerfi, Laszlo ^{[1
]}

Wegkamp, Marten ^{[2
]}

机构：

[1] Budapest Univ Technol & Econ, Dept Comp Sci & Informat Theory, H-1117 Budapest, Hungary

[2] Florida State Univ, Dept Stat, Tallahassee, FL 32306 USA

来源：

IEEE TRANSACTIONS ON INFORMATION THEORY | 2008年 / 54卷 / 02期

基金：

美国国家科学基金会;

关键词：

regression estimation with restriction; least squares estimates; vector quantization; finite-sample bounds;

D O I：

10.1109/TIT.2007.913565

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

The authors discuss quantization or clustering of non-parametric 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.

引用

页码：867 / 874

页数：8

共 10 条

[1]

Anthony M., 1999, Neural Network Learning: Theoretical Foundations, V9

[2] Short description of an alternative simplified method for screening recombinant clones within the "AdEasy-System" by Duplex-PCR -: art. no. 1 [J].

Antolovic, D ;

Koch, M ;

Bohlmann, I ;

Kienle, P ;

Büchler, M ;

Weitz, J .

BMC BIOTECHNOLOGY, 2005, 5 (1)

[3]

DEVROYE L, 1983, P 4 PANN S MATH STAT, P67

[4]

DEVROYE L, 2001, COMBINATIONAL METHOD

[5]

GYORFI L, 1991, NATO ADV SCI I C-MAT, V335, P329

[6]

Gyorfi L., 2002, A Distribution-Free Theory of Nonparametric Regression, DOI DOI 10.1007/B97848

[7]

Lee WS, 1996, IEEE T INFORM THEORY, V42, P2118, DOI 10.1109/18.556601

[8] Empirical quantizer design in the presence of source noise or channel noise [J].

Linder, T ;

Lugosi, G ;

Zeger, K .

IEEE TRANSACTIONS ON INFORMATION THEORY, 1997, 43 (02) :612-623

[9]

POLLARD DP, 1982, ANN STAT, V16, P919

[10]

WEGKAMP MH, 1999, ENTROPY METHODS STAT, V25

← 1 →