摘要
在水库径流预测中应用模糊聚类分析方法优选网络结构 ,优化后的网络具有较高的效率 ,并取得了较好的拟合。
Feedforward neural network has been used widely in water conservancy and hydropower engineering. The difficulty in application is how to choose the structure of networks. In this paper, fuzzy cluster method is used to determine the number of hidden nodes. The optimal network is successfully applied in runoff forecast, it is more efficient and precise.
出处
《水电能源科学》
2002年第3期35-37,共3页
Water Resources and Power
关键词
径流预测
隐层节点
模糊聚类
前馈神经网络
runoff forecast
hidden nodes
extraction of feature
fuzzy cluster
feedforward neural networks
