Neural network structure for modeling of a complex multivariable function based on data samples

A problem of approximating a multivariable function measured against the background of additive noise and characterized by qualitatively different dynamic properties in separate subdomains of definition is considered. The solution of the problem using a specialized modular neural network structure L...

Full description

Saved in:
Bibliographic Details
Published inJournal of computer & systems sciences international Vol. 46; no. 1; pp. 68 - 77
Main Authors Gavrilova, E. V., Mishulina, O. A., Shcherbinina, M. V.
Format Journal Article
LanguageEnglish
Published Silver Spring Springer Nature B.V 01.02.2007
Subjects
Neural networks
Studies
Online AccessGet full text

Cover

More Information
Summary:A problem of approximating a multivariable function measured against the background of additive noise and characterized by qualitatively different dynamic properties in separate subdomains of definition is considered. The solution of the problem using a specialized modular neural network structure LINA is proposed, and rules for its training and operation are given. The solution of the navigation problem for an aircraft based on the measurement of distances to radio beacons is given as an example of application of the developed neural network technique.[PUBLICATION ABSTRACT]
ISSN:1064-2307
1555-6530
DOI:10.1134/S106423070701008X