摘要： 文中提出一个加快BP型人工神经元网络学习速度的方法,它包括两个方面的内容,一是将归一化后用来输入BP网络的模式特征向量进行编码,二是将BP网络输入层的节点数加以扩充,从而使每一个码位对应一个输入节点。理论分析和实验结果均表明,该方法可以有效地缩短BP网络的学习时间。

关键词: BP网络, 特征值编码, 节点扩充, 快速学习

Abstract: In this paper, an approach to speed up the learning of BP neural network is described. The approach consists of two methods.One is coding the normalized value of pattern features. Another method is expanding the input layer nodes of BP neural network.Every bit of coded features is fed in a node of expanded input layer.The experimental results show that this approach can speed up the learning of BP neural network effectively.

中图分类号: 

• TP183