A Kalman filtering algorithm for regularization networks

Author

De Nicolao, G. ; Ferrari-Trecate, G.

Author_Institution

Dipartimento di Inf. e Sistemistica, Pavia Univ., Italy

Volume

4

fYear

2000

fDate

2000

Firstpage

2220

Abstract

Regularization networks are nonparametric estimators obtained from the application of Tychonov regularization or Bayes estimation to the hypersurface reconstruction problem. With the usual algorithm, the computation of the weights scales as O(n³) where n is the number of data. We show that for a class of monodimensional problems, the complexity can be reduced to O(n) by a suitable algorithm based on spectral factorization and Kalman filtering. The procedure applies also to smoothing splines and, in a multidimensional context, to additive regularization networks

Keywords

Bayes methods; Kalman filters; computational complexity; filtering theory; nonparametric statistics; Bayes estimation; Kalman filtering algorithm; Tychonov regularization; hypersurface reconstruction problem; monodimensional problems; nonparametric estimators; regularization networks; spectral factorization; Cost function; Electronic mail; Filtering algorithms; H infinity control; Kalman filters; Measurement errors; Neural networks; Neurons; Smoothing methods; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, 2000. Proceedings of the 2000

Conference_Location

Chicago, IL

ISSN

0743-1619

Print_ISBN

0-7803-5519-9

Type

conf

DOI

10.1109/ACC.2000.878574

Filename

878574