一种改进的BP算法

journal6 ›› 2003, Vol. 24 ›› Issue (4): 79-81.

一种改进的BP算法

(吉首大学预科部，湖南吉首 416000)

出版日期:2003-12-15 发布日期:2012-11-06
作者简介:张清良(1965-)，女(回族)，湖南省邵阳市人，吉首大学预科部讲师，主要从事计算数学研究.

Improvement of the BP Algorithm

(Department of Prepartory Science,Jishou University,Jishou 416000,Hunan China)

Online:2003-12-15 Published:2012-11-06

摘要/Abstract

摘要：传统的BP算法收敛速度慢，利用附加动量因子和自适应学习速率改进了传统的BP算法，它对于BP网络结构优化，提高收敛速度有明显效果.

关键词： BP算法, 动量因子, 自适应学习速率

Abstract: The BP algorithm which belongs to the neural net model is the most extensive study algorthm at present.The convergence speed of traditiond BP algorithm is slow.In oder to improve the traditiond BP algorithm,this paper is trying to use the power divisor and the adaptive study rate.And it will have obvious efficency on the adjustment of BP net structure and improving the speed of convergence.

Key words: the BP algorithm, the power divisor, the adaptive study rate

张清良. 一种改进的BP算法[J]. journal6, 2003, 24(4): 79-81.

ZHANG Qing-Liang. Improvement of the BP Algorithm[J]. journal6, 2003, 24(4): 79-81.

[1]	温芝元, 曹乐平. 神经网络在项目财务评价上的应用[J]. journal6, 2005, 26(3): 86-90.
[2]	何国光, 周坚强. 基于前向神经网络的知识获取[J]. journal6, 2002, 23(2): 62-65.
[3]	胡师彦. 混合遗传BP算法在图象识别中的应用[J]. journal6, 2001, 22(4): 57-59.