Designer networks for time series processing

C Svarer; Lars Kai Hansen; Jan Larsen; Carl Edward Rasmussen

doi:10.1109/NNSP.1993.471881

Designer networks for time series processing

C Svarer, Lars Kai Hansen, Jan Larsen, Carl Edward Rasmussen

Technical University of Denmark

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

285 Downloads (Pure)

Abstract

The conventional tapped-delay neural net may be analyzed using statistical methods and the results of such analysis can be applied to model optimization. The authors review and extend efforts to demonstrate the power of this strategy within time series processing. They attempt to design compact networks using the so-called optima brain damage (OBD) method. The benefits from compact architectures are three-fold. Their generalization ability is at least comparable,they involve less computational burden, and they are faster to adapt if the environment changes. It is shown that the generalization error of the network may be estimated, without extensive cross-validation, using a modification of Akaike's final prediction error (FPE) estimate (1969)

Original language	English
Title of host publication	Proceedings of the IEEE-SP Workshop on Neural Networks for Signal Processing
Publisher	IEEE
Publication date	1993
Pages	78-87
ISBN (Print)	07-80-30928-6
DOIs	https://doi.org/10.1109/NNSP.1993.471881
Publication status	Published - 1993
Event	1993 IEEE Workshop on Neural Networks for Signal Processing - , United States Duration: 6 Sept 1993 → 9 Sept 1993 Conference number: 3 https://ieeexplore.ieee.org/xpl/conhome/3312/proceeding

Conference

Conference	1993 IEEE Workshop on Neural Networks for Signal Processing
Number	3
Country/Territory	United States
Period	06/09/1993 → 09/09/1993
Internet address	https://ieeexplore.ieee.org/xpl/conhome/3312/proceeding

Bibliographical note

Copyright: 1993 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.1993.471881

SvarerFinal published version, 545 KB

OpenUrl availability

Full text

Cite this

@inproceedings{e26b205d6fab4a39904e9941e400786d,

title = "Designer networks for time series processing",

abstract = "The conventional tapped-delay neural net may be analyzed using statistical methods and the results of such analysis can be applied to model optimization. The authors review and extend efforts to demonstrate the power of this strategy within time series processing. They attempt to design compact networks using the so-called optima brain damage (OBD) method. The benefits from compact architectures are three-fold. Their generalization ability is at least comparable,they involve less computational burden, and they are faster to adapt if the environment changes. It is shown that the generalization error of the network may be estimated, without extensive cross-validation, using a modification of Akaike's final prediction error (FPE) estimate (1969)",

author = "C Svarer and Hansen, {Lars Kai} and Jan Larsen and Rasmussen, {Carl Edward}",

note = "Copyright: 1993 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; 1993 IEEE Workshop on Neural Networks for Signal Processing, NNSP 1993 ; Conference date: 06-09-1993 Through 09-09-1993",

year = "1993",

doi = "10.1109/NNSP.1993.471881",

language = "English",

isbn = "07-80-30928-6",

pages = "78--87",

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

publisher = "IEEE",

address = "United States",

url = "https://ieeexplore.ieee.org/xpl/conhome/3312/proceeding",

}

TY - GEN

T1 - Designer networks for time series processing

AU - Svarer, C

AU - Hansen, Lars Kai

AU - Larsen, Jan

AU - Rasmussen, Carl Edward

N1 - Conference code: 3

PY - 1993

Y1 - 1993

N2 - The conventional tapped-delay neural net may be analyzed using statistical methods and the results of such analysis can be applied to model optimization. The authors review and extend efforts to demonstrate the power of this strategy within time series processing. They attempt to design compact networks using the so-called optima brain damage (OBD) method. The benefits from compact architectures are three-fold. Their generalization ability is at least comparable,they involve less computational burden, and they are faster to adapt if the environment changes. It is shown that the generalization error of the network may be estimated, without extensive cross-validation, using a modification of Akaike's final prediction error (FPE) estimate (1969)

AB - The conventional tapped-delay neural net may be analyzed using statistical methods and the results of such analysis can be applied to model optimization. The authors review and extend efforts to demonstrate the power of this strategy within time series processing. They attempt to design compact networks using the so-called optima brain damage (OBD) method. The benefits from compact architectures are three-fold. Their generalization ability is at least comparable,they involve less computational burden, and they are faster to adapt if the environment changes. It is shown that the generalization error of the network may be estimated, without extensive cross-validation, using a modification of Akaike's final prediction error (FPE) estimate (1969)

U2 - 10.1109/NNSP.1993.471881

DO - 10.1109/NNSP.1993.471881

M3 - Article in proceedings

SN - 07-80-30928-6

SP - 78

EP - 87

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

PB - IEEE

T2 - 1993 IEEE Workshop on Neural Networks for Signal Processing

Y2 - 6 September 1993 through 9 September 1993

ER -

Designer networks for time series processing

Abstract

Conference

Bibliographical note

Access to Document

OpenUrl availability

Fingerprint

Cite this