Measuring Order Parameters Based on Neural Network Technologies

We show convergence of a procedure for multiple training of a neural network, when during multiple repetitions of the solution of a binary classification task based on unchanged packets of training samples the neural emulator can initially demonstrate a chaotic set of values for the feature weights,...

Full description

Saved in:

Bibliographic Details
Published in:	Measurement techniques Vol. 58; no. 4; pp. 462 - 466
Main Authors:	Vokhmina, Yu. V., Eskov, V. M., Gavrilenko, T. V., Filatova, O. E.
Format:	Journal Article
Language:	English
Published:	New York Springer US 01-07-2015 Springer Nature B.V
Subjects:	Analytical Chemistry Chaos theory Characterization and Evaluation of Materials Classification Convergence International Measurement Science and Instrumentation Measurement techniques Medical and Biological Measurements Medicine Nervous system Neural networks Order parameters Patients Physical Chemistry Physics Physics and Astronomy Repetition Studies Tasks Training Weight reduction Russia biosystem theory of chaos and self-organization neural network order parameters
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	We show convergence of a procedure for multiple training of a neural network, when during multiple repetitions of the solution of a binary classification task based on unchanged packets of training samples the neural emulator can initially demonstrate a chaotic set of values for the feature weights, which as the number of iterations increases is reduced to a certain ranked series.
Bibliography:	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23
ISSN:	0543-1972 1573-8906
DOI:	10.1007/s11018-015-0735-x