Please use this identifier to cite or link to this item: http://ea.donntu.edu.ua:8080/jspui/handle/123456789/15951
Title: ПРОГРАММНАЯ РЕАЛИЗАЦИЯ НЕЙРОННЫХ СЕТЕЙ ДЛЯ ОБУЧЕНИЯ В СИСТЕМАХ РЕАЛЬНОГО ВРЕМЕНИ
Other Titles: Program realization of neural networks for learning in real-time systems
Authors: Скобцов, Ю. А.
Ткаченко, А. В.
Skobtsov, Y.A.
Tkachenko, A. V.
Keywords: Mathlab
neural networks
real-time systems
системы реального времени
нейронные сети
Issue Date: 2005
Publisher: ДонНТУ
Citation: Наукові праці Донецького національного технічного університету. Серія: “Обчислювальна техніка та автоматизація”. Випуск 90 — Донецьк: ДонНТУ, 2005
Description: Problem of program realization of artificial neural networks for real-time system is considered. Best methods for neural network training are overviewed. Resilient backpropagation algorithm is considered as one of the most efficient weight adaptation method. Application of conjugate graph for gradient generation is shown. Problem of weight initialization is considered and Ngueyn-Widrow method is applied to give good first approximation. Program library implementing overviewed methods is developed on С High speed of the library is shown in comparison with Mathlab.
URI: http://ea.donntu.edu.ua/handle/123456789/15951
Appears in Collections:Випуск 90

Files in This Item:
File Description SizeFormat 
119-126.pdf5,4 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.