摘要
提出了一种改善ＢＰ网络学习能力的样本添加法，它是根据人类学习知识时存在遗忘现象而设计的．首先阐明其工作原理与学习算法；接着将它用于ＢＰ网络之中，以处理三倍冗余的表决问题．结果表明，该法的应用可以明显地提高ＢＰ网络的学习能力，而且，单一样本比成批样本的添加效果更好．
A sample increasing method for BP neural networks is proposed. It is designed based on the phenomenon of forgetting when men are learning knowledge. After the principle and algorithm of the method are stated, the BP neural network is applied to voting processing. The results show the good performance of the method for convergence of BP neural nets.
出处
《大连理工大学学报》
EI
CAS
CSCD
北大核心
1997年第5期564-567,共4页
Journal of Dalian University of Technology
关键词
神经网络
算法
BP网络
样本添加法
学习能力
neural networks
algorithms/BP networks
voting
sample increasing method
