Estimation of error confidence intervals for the regression of real-valued functions

Rhee Man Kil, Imhoi Koo

Research output: Contribution to conferencePaper

1 Citation (Scopus)

Abstract

This paper presents a new method of estimating the error confidence interval defined by the absolute value of difference between the true (or general) and empirical risks for the regression of real-valued functions. The theoretical bounds of error confidence intervals can be derived in the sense of probably approximately correct (PAC) learning. However, these theoretical bounds are too overestimated and not well fitted to the empirical data. In this sense, a new estimation model of error confidence intervals which can explain the behavior of general error more faithfully to the given samples, is suggested. To show the validity of our model, the error confidence intervals for the approximation of 2-D function and the prediction of Mackey-Glass time series, are estimated and compared with the experimental results.

Original languageEnglish (US)
Pages1002-1006
Number of pages5
StatePublished - Jan 1 2002
Event2002 International Joint Conference on Neural Networks (IJCNN '02) - Honolulu, HI, United States
Duration: May 12 2002May 17 2002

Conference

Conference2002 International Joint Conference on Neural Networks (IJCNN '02)
CountryUnited States
CityHonolulu, HI
Period5/12/025/17/02

Fingerprint

Time series
Glass

All Science Journal Classification (ASJC) codes

  • Software
  • Artificial Intelligence

Cite this

Kil, R. M., & Koo, I. (2002). Estimation of error confidence intervals for the regression of real-valued functions. 1002-1006. Paper presented at 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, United States.
Kil, Rhee Man ; Koo, Imhoi. / Estimation of error confidence intervals for the regression of real-valued functions. Paper presented at 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, United States.5 p.
@conference{24c50e3fa17d45ec95ae292b69243dd1,
title = "Estimation of error confidence intervals for the regression of real-valued functions",
abstract = "This paper presents a new method of estimating the error confidence interval defined by the absolute value of difference between the true (or general) and empirical risks for the regression of real-valued functions. The theoretical bounds of error confidence intervals can be derived in the sense of probably approximately correct (PAC) learning. However, these theoretical bounds are too overestimated and not well fitted to the empirical data. In this sense, a new estimation model of error confidence intervals which can explain the behavior of general error more faithfully to the given samples, is suggested. To show the validity of our model, the error confidence intervals for the approximation of 2-D function and the prediction of Mackey-Glass time series, are estimated and compared with the experimental results.",
author = "Kil, {Rhee Man} and Imhoi Koo",
year = "2002",
month = "1",
day = "1",
language = "English (US)",
pages = "1002--1006",
note = "2002 International Joint Conference on Neural Networks (IJCNN '02) ; Conference date: 12-05-2002 Through 17-05-2002",

}

Kil, RM & Koo, I 2002, 'Estimation of error confidence intervals for the regression of real-valued functions', Paper presented at 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, United States, 5/12/02 - 5/17/02 pp. 1002-1006.

Estimation of error confidence intervals for the regression of real-valued functions. / Kil, Rhee Man; Koo, Imhoi.

2002. 1002-1006 Paper presented at 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, United States.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Estimation of error confidence intervals for the regression of real-valued functions

AU - Kil, Rhee Man

AU - Koo, Imhoi

PY - 2002/1/1

Y1 - 2002/1/1

N2 - This paper presents a new method of estimating the error confidence interval defined by the absolute value of difference between the true (or general) and empirical risks for the regression of real-valued functions. The theoretical bounds of error confidence intervals can be derived in the sense of probably approximately correct (PAC) learning. However, these theoretical bounds are too overestimated and not well fitted to the empirical data. In this sense, a new estimation model of error confidence intervals which can explain the behavior of general error more faithfully to the given samples, is suggested. To show the validity of our model, the error confidence intervals for the approximation of 2-D function and the prediction of Mackey-Glass time series, are estimated and compared with the experimental results.

AB - This paper presents a new method of estimating the error confidence interval defined by the absolute value of difference between the true (or general) and empirical risks for the regression of real-valued functions. The theoretical bounds of error confidence intervals can be derived in the sense of probably approximately correct (PAC) learning. However, these theoretical bounds are too overestimated and not well fitted to the empirical data. In this sense, a new estimation model of error confidence intervals which can explain the behavior of general error more faithfully to the given samples, is suggested. To show the validity of our model, the error confidence intervals for the approximation of 2-D function and the prediction of Mackey-Glass time series, are estimated and compared with the experimental results.

UR - http://www.scopus.com/inward/record.url?scp=0036078730&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036078730&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0036078730

SP - 1002

EP - 1006

ER -

Kil RM, Koo I. Estimation of error confidence intervals for the regression of real-valued functions. 2002. Paper presented at 2002 International Joint Conference on Neural Networks (IJCNN '02), Honolulu, HI, United States.