[1]孟斌,冯永杰,翟玉庆.前馈神经网络中BP算法的一种改进[J].东南大学学报(自然科学版),2001,31(4):40-42.[doi:10.3969/j.issn.1001-0505.2001.04.010]
 Meng Bin,Feng Yongjie,Zhai Yuqing.A Modified Algorithm for Feedforward Neural Networks[J].Journal of Southeast University (Natural Science Edition),2001,31(4):40-42.[doi:10.3969/j.issn.1001-0505.2001.04.010]
点击复制

前馈神经网络中BP算法的一种改进()
分享到:

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
31
期数:
2001年第4期
页码:
40-42
栏目:
自动化
出版日期:
2001-07-20

文章信息/Info

Title:
A Modified Algorithm for Feedforward Neural Networks
作者:
孟斌 冯永杰 翟玉庆
东南大学计算机科学与工程系, 南京 210096
Author(s):
Meng Bin Feng Yongjie Zhai Yuqing
Department of Computer Science and Technology, Southeast University, Nanjing 210096, China)
关键词:
前馈神经网络 BP学习算法 收敛速度 学习步长
Keywords:
feedforward neural networks BP learning algorithm convergence speed learning step
分类号:
TP183
DOI:
10.3969/j.issn.1001-0505.2001.04.010
摘要:
在传统的BP算法基础上,提出了一种改进的BP学习算法.先加入描述网络复杂性的量,使算法能够考虑到网络的连接复杂性,进而有可能删除掉冗余的连接甚至节点; 接着提出对网络的学习步长的动态调整,以此来尽量避免传统学习中的学习速度过慢和反复震荡; 然后给出新的算法是高阶非线性收敛的证明; 最后通过实验说明新的BP算法在一定程度上可减少网络的复杂性,有着比传统算法更快的收敛速度.
Abstract:
Based on conventional BP algorithm, a modified BP learning algorithm is proposed. A factor that is used to describe the ANN’s complexity is added in order to evaluate the complexity. This makes it possible that the redundant connections can be decreased, and the redundant neurons can be nullified even. Then, a method that can dynamically adjust the learning step is presented. So the learning step can be speeded and the sway phenomenon can be minimized.Next, it is proved that the new algorithm is super-linearly convergent. Finally, simulation results illustrate that the algorithm can decrease ANN’s complexity with higher convergence speed under certain conditions than conventional BP algorithm.

参考文献/References:

[1] 胡守仁.神经网络导论.长沙:国防科技大学出版社,1998.113~121
[2] 胡守仁.神经网络应用技术.长沙:国防科技大学出版社,1998.82~86
[3] Rumellhart D E,Hinton G E,Williams R J.Learning internal representations by error propagation.Nature,1986,323(6188):533~536
[4] 罗四维.人工神经网络建造.北京:中国铁道出版社,1998.100~102

备注/Memo

备注/Memo:
作者简介:孟斌,男,1975年生,硕士研究生.
更新日期/Last Update: 2001-07-20