A neural architecture for nonlinear adaptive filtering of time series

Nils Hoffmann; Jan Larsen

doi:10.1109/NNSP.1991.239488

A neural architecture for nonlinear adaptive filtering of time series

Nils Hoffmann, Jan Larsen

Technical University of Denmark

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

464 Downloads (Pure)

Abstract

A neural architecture for adaptive filtering which incorporates a modularization principle is proposed. It facilitates a sparse parameterization, i.e. fewer parameters have to be estimated in a supervised training procedure. The main idea is to use a preprocessor which determines the dimension of the input space and can be designed independently of the subsequent nonlinearity. Two suggestions for the preprocessor are presented: the derivative preprocessor and the principal component analysis. A novel implementation of fixed Volterra nonlinearities is given. It forces the boundedness of the polynominals by scaling and limiting the inputs signals. The nonlinearity is constructed from Chebychev polynominals. The authors apply a second-order algorithm for updating the weights for adaptive nonlinearities. Finally the simulations indicate that the two kinds of preprocessing tend to complement each other while there is no obvious difference between the performance of the ANL and FNL

Original language	English
Title of host publication	Proceedings of the IEEE Workshop Neural Networks for Signal Processing
Publisher	IEEE
Publication date	1991
Pages	533-542
ISBN (Print)	0-7803-0118-8
DOIs	https://doi.org/10.1109/NNSP.1991.239488
Publication status	Published - 1991
Event	1991 IEEE Workshop on Neural Networks for Signal Processing - Princeton, United States Duration: 30 Sept 1991 → 1 Oct 1991 Conference number: 1 https://ieeexplore.ieee.org/xpl/conhome/574/proceeding

Conference

Conference	1991 IEEE Workshop on Neural Networks for Signal Processing
Number	1
Country/Territory	United States
City	Princeton
Period	30/09/1991 → 01/10/1991
Internet address	https://ieeexplore.ieee.org/xpl/conhome/574/proceeding

Bibliographical note

Copyright: 1991 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.1991.239488

HoffmannFinal published version, 538 KB

OpenUrl availability

Full text

Cite this

@inproceedings{6777128c59fa47789df38acc101d1f6a,

title = "A neural architecture for nonlinear adaptive filtering of time series",

abstract = "A neural architecture for adaptive filtering which incorporates a modularization principle is proposed. It facilitates a sparse parameterization, i.e. fewer parameters have to be estimated in a supervised training procedure. The main idea is to use a preprocessor which determines the dimension of the input space and can be designed independently of the subsequent nonlinearity. Two suggestions for the preprocessor are presented: the derivative preprocessor and the principal component analysis. A novel implementation of fixed Volterra nonlinearities is given. It forces the boundedness of the polynominals by scaling and limiting the inputs signals. The nonlinearity is constructed from Chebychev polynominals. The authors apply a second-order algorithm for updating the weights for adaptive nonlinearities. Finally the simulations indicate that the two kinds of preprocessing tend to complement each other while there is no obvious difference between the performance of the ANL and FNL",

author = "Nils Hoffmann and Jan Larsen",

note = "Copyright: 1991 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; 1991 IEEE Workshop on Neural Networks for Signal Processing, NNSP 1991 ; Conference date: 30-09-1991 Through 01-10-1991",

year = "1991",

doi = "10.1109/NNSP.1991.239488",

language = "English",

isbn = "0-7803-0118-8",

pages = "533--542",

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

publisher = "IEEE",

address = "United States",

url = "https://ieeexplore.ieee.org/xpl/conhome/574/proceeding",

}

TY - GEN

T1 - A neural architecture for nonlinear adaptive filtering of time series

AU - Hoffmann, Nils

AU - Larsen, Jan

N1 - Conference code: 1

PY - 1991

Y1 - 1991

N2 - A neural architecture for adaptive filtering which incorporates a modularization principle is proposed. It facilitates a sparse parameterization, i.e. fewer parameters have to be estimated in a supervised training procedure. The main idea is to use a preprocessor which determines the dimension of the input space and can be designed independently of the subsequent nonlinearity. Two suggestions for the preprocessor are presented: the derivative preprocessor and the principal component analysis. A novel implementation of fixed Volterra nonlinearities is given. It forces the boundedness of the polynominals by scaling and limiting the inputs signals. The nonlinearity is constructed from Chebychev polynominals. The authors apply a second-order algorithm for updating the weights for adaptive nonlinearities. Finally the simulations indicate that the two kinds of preprocessing tend to complement each other while there is no obvious difference between the performance of the ANL and FNL

AB - A neural architecture for adaptive filtering which incorporates a modularization principle is proposed. It facilitates a sparse parameterization, i.e. fewer parameters have to be estimated in a supervised training procedure. The main idea is to use a preprocessor which determines the dimension of the input space and can be designed independently of the subsequent nonlinearity. Two suggestions for the preprocessor are presented: the derivative preprocessor and the principal component analysis. A novel implementation of fixed Volterra nonlinearities is given. It forces the boundedness of the polynominals by scaling and limiting the inputs signals. The nonlinearity is constructed from Chebychev polynominals. The authors apply a second-order algorithm for updating the weights for adaptive nonlinearities. Finally the simulations indicate that the two kinds of preprocessing tend to complement each other while there is no obvious difference between the performance of the ANL and FNL

U2 - 10.1109/NNSP.1991.239488

DO - 10.1109/NNSP.1991.239488

M3 - Article in proceedings

SN - 0-7803-0118-8

SP - 533

EP - 542

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

PB - IEEE

T2 - 1991 IEEE Workshop on Neural Networks for Signal Processing

Y2 - 30 September 1991 through 1 October 1991

ER -

A neural architecture for nonlinear adaptive filtering of time series

Abstract

Conference

Bibliographical note

Access to Document

OpenUrl availability

Fingerprint

Cite this