Approximately Unbiased Estimation of Conditional Variance in Heteroscedastic Kernel Ridge Regression

Gavin C. Cawley; Nicola L. C. Talbot; Robert J. Foxall; Stephen R. Dorling; Danilo P. Mandic

Approximately Unbiased Estimation of Conditional Variance in Heteroscedastic Kernel Ridge Regression

Gavin C. Cawley, Nicola L. C. Talbot, Robert J. Foxall, Stephen R. Dorling, Danilo P. Mandic

Research output: Contribution to conference › Paper

Abstract

In this paper we extend a form of kernel ridge regression for data characterised by a heteroscedastic noise process (introduced in Foxall et al. [1]) in order to provide approximately unbiased estimates of the conditional variance of the target distribution. This is achieved by the use of the leave-one-out cross-validation estimate of the conditional mean when fitting the model of the conditional variance. The elimination of this bias is demonstrated on synthetic dataset where the true conditional variance is known.

Original language	English
Pages	209-214
Number of pages	6
Publication status	Published - Apr 2003
Event	European Symposium on Artificial Neural Networks - Bruges, Belgium Duration: 23 Apr 2003 → 25 Apr 2003

Conference

Conference	European Symposium on Artificial Neural Networks
Abbreviated title	ESANN-2003
Country/Territory	Belgium
City	Bruges
Period	23/04/03 → 25/04/03

Cite this

@conference{e09988ac6078425fa948dbec6571a113,

title = "Approximately Unbiased Estimation of Conditional Variance in Heteroscedastic Kernel Ridge Regression",

abstract = "In this paper we extend a form of kernel ridge regression for data characterised by a heteroscedastic noise process (introduced in Foxall et al. [1]) in order to provide approximately unbiased estimates of the conditional variance of the target distribution. This is achieved by the use of the leave-one-out cross-validation estimate of the conditional mean when fitting the model of the conditional variance. The elimination of this bias is demonstrated on synthetic dataset where the true conditional variance is known.",

author = "Cawley, {Gavin C.} and Talbot, {Nicola L. C.} and Foxall, {Robert J.} and Dorling, {Stephen R.} and Mandic, {Danilo P.}",

year = "2003",

month = apr,

language = "English",

pages = "209--214",

note = "European Symposium on Artificial Neural Networks, ESANN-2003 ; Conference date: 23-04-2003 Through 25-04-2003",

}

TY - CONF

T1 - Approximately Unbiased Estimation of Conditional Variance in Heteroscedastic Kernel Ridge Regression

AU - Cawley, Gavin C.

AU - Talbot, Nicola L. C.

AU - Foxall, Robert J.

AU - Dorling, Stephen R.

AU - Mandic, Danilo P.

PY - 2003/4

Y1 - 2003/4

N2 - In this paper we extend a form of kernel ridge regression for data characterised by a heteroscedastic noise process (introduced in Foxall et al. [1]) in order to provide approximately unbiased estimates of the conditional variance of the target distribution. This is achieved by the use of the leave-one-out cross-validation estimate of the conditional mean when fitting the model of the conditional variance. The elimination of this bias is demonstrated on synthetic dataset where the true conditional variance is known.

AB - In this paper we extend a form of kernel ridge regression for data characterised by a heteroscedastic noise process (introduced in Foxall et al. [1]) in order to provide approximately unbiased estimates of the conditional variance of the target distribution. This is achieved by the use of the leave-one-out cross-validation estimate of the conditional mean when fitting the model of the conditional variance. The elimination of this bias is demonstrated on synthetic dataset where the true conditional variance is known.

M3 - Paper

SP - 209

EP - 214

T2 - European Symposium on Artificial Neural Networks

Y2 - 23 April 2003 through 25 April 2003

ER -