Variable projections neural network training

The training of some types of neural networks leads to separable non-linear least squares problems. These problems may be ill-conditioned and require special techniques. A robust algorithm based on the Variable Projections method of Golub and Pereyra is designed for a class of feed-forward neural networks and tested on benchmark examples and real data.

MoreLess

Year of publication:	2006
Authors:	Pereyra, V. ; Scherer, G. ; Wong, F.
Published in:	Mathematics and Computers in Simulation (MATCOM). - Elsevier, ISSN 0378-4754. - Vol. 73.2006, 1, p. 231-243
Publisher:	Elsevier
Subject:	Neural networks \| Variable projection algorithm

More details

Type of publication:	Article
Source:	RePEc - Research Papers in Economics

Persistent link: https://www.econbiz.de/10010749923