摘要
误差反向传播算法(BP算法)是训练多层网络的常用算法,但难于确定网络的有效隐节点数,本文提出了一个有效的算法来自动优选隐层节点数,文中用龙羊峡水电站历年10月份的系列实测资料进行了实例研究,取得了初步令人满意的结果。
BP algorithm is probably the most successful training procedure for multi-layer neural nets, but it is difficult to determine the optimal number of hidden units. This paper give a new method to choose the number of hidden node. The result of application in Long Yangxia Hydropower Station is satisfactory.
出处
《应用基础与工程科学学报》
EI
CSCD
1995年第2期99-104,共6页
Journal of Basic Science and Engineering
关键词
神经网络
BP算法
隐节点数
径流预报
neural network, BP algorithm, hidden node, runoff forecast
