A generalization error estimate for nonlinear systems

Jan Larsen

doi:10.1109/NNSP.1992.253710

A generalization error estimate for nonlinear systems

Jan Larsen

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

384 Downloads (Pure)

Abstract

A new estimate (GEN) of the generalization error is presented. The estimator is valid for both incomplete and nonlinear models. An incomplete model is characterized in that it does not model the actual nonlinear relationship perfectly. The GEN estimator has been evaluated by simulating incomplete models of linear and simple neural network systems. Within the linear system GEN is compared to the final prediction error criterion and the leave-one-out cross-validation technique. It was found that the GEN estimate of the true generalization error is less biased on the average. It is concluded that GEN is an applicable alternative in estimating the generalization at the expense of an increased complexity

Original language	English
Title of host publication	Proceedings of the IEEE-SP Workshop Neural Networks for Signal Processing
Publisher	IEEE
Publication date	1992
Pages	29-38
ISBN (Print)	07-80-30557-4
DOIs	https://doi.org/10.1109/NNSP.1992.253710
Publication status	Published - 1992
Event	1992 IEEE-SP Workshop of Neural Networks for Signal Processing - Helsingør, Denmark Duration: 31 Aug 1992 → 2 Sep 1992 http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=631

Workshop

Workshop	1992 IEEE-SP Workshop of Neural Networks for Signal Processing
Country	Denmark
City	Helsingør
Period	31/08/1992 → 02/09/1992
Internet address	http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=631

Bibliographical note

Copyright: 1992 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE

Access to Document

10.1109/NNSP.1992.253710

LarsenFinal published version, 455 KB

Fingerprint Dive into the research topics of 'A generalization error estimate for nonlinear systems'. Together they form a unique fingerprint.

Cite this

@inproceedings{718ba01530bb4f32b8504c8c464cc0d4,

title = "A generalization error estimate for nonlinear systems",

abstract = "A new estimate (GEN) of the generalization error is presented. The estimator is valid for both incomplete and nonlinear models. An incomplete model is characterized in that it does not model the actual nonlinear relationship perfectly. The GEN estimator has been evaluated by simulating incomplete models of linear and simple neural network systems. Within the linear system GEN is compared to the final prediction error criterion and the leave-one-out cross-validation technique. It was found that the GEN estimate of the true generalization error is less biased on the average. It is concluded that GEN is an applicable alternative in estimating the generalization at the expense of an increased complexity",

author = "Jan Larsen",

note = "Copyright: 1992 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE; 1992 IEEE-SP Workshop of Neural Networks for Signal Processing ; Conference date: 31-08-1992 Through 02-09-1992",

year = "1992",

doi = "10.1109/NNSP.1992.253710",

language = "English",

isbn = "07-80-30557-4",

pages = "29--38",

booktitle = "Proceedings of the IEEE-SP Workshop Neural Networks for Signal Processing",

publisher = "IEEE",

address = "United States",

url = "http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=631",

}

TY - GEN

T1 - A generalization error estimate for nonlinear systems

AU - Larsen, Jan

N1 - Copyright: 1992 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE

PY - 1992

Y1 - 1992

N2 - A new estimate (GEN) of the generalization error is presented. The estimator is valid for both incomplete and nonlinear models. An incomplete model is characterized in that it does not model the actual nonlinear relationship perfectly. The GEN estimator has been evaluated by simulating incomplete models of linear and simple neural network systems. Within the linear system GEN is compared to the final prediction error criterion and the leave-one-out cross-validation technique. It was found that the GEN estimate of the true generalization error is less biased on the average. It is concluded that GEN is an applicable alternative in estimating the generalization at the expense of an increased complexity

AB - A new estimate (GEN) of the generalization error is presented. The estimator is valid for both incomplete and nonlinear models. An incomplete model is characterized in that it does not model the actual nonlinear relationship perfectly. The GEN estimator has been evaluated by simulating incomplete models of linear and simple neural network systems. Within the linear system GEN is compared to the final prediction error criterion and the leave-one-out cross-validation technique. It was found that the GEN estimate of the true generalization error is less biased on the average. It is concluded that GEN is an applicable alternative in estimating the generalization at the expense of an increased complexity

U2 - 10.1109/NNSP.1992.253710

DO - 10.1109/NNSP.1992.253710

M3 - Article in proceedings

SN - 07-80-30557-4

SP - 29

EP - 38

BT - Proceedings of the IEEE-SP Workshop Neural Networks for Signal Processing

PB - IEEE

T2 - 1992 IEEE-SP Workshop of Neural Networks for Signal Processing

Y2 - 31 August 1992 through 2 September 1992

ER -