摘要
介绍了径向基函数(RBF)神经网络的结构、原理和训练算法。以某市为研究对象,建立了RBF神经网络工业取水量预测模型,采用最近邻聚类学习算法确定径向基函数的宽度、聚类中心和权值。结果表明:RBF模型具有较强的非线性处理能力和逼近能力,且结构简洁、学习速度快、预测精度高,泛化能力强,克服了BP神经网络学习过程收敛过分依赖于初值和可能出现局部收敛的缺陷。
This paper introduces the structure,principle and training algorithm of RBF Neural Network,taking a city as research subject,sets up the model of water demand forecasting for industry based on RBF Neural Network,and determines the width,clustering center and weight of RBF by the nearest neighbor clustering algorithm.The results showed that the RBF Neural Network has strong nonlinear processing ability and approximation capability,moreover it has simple structure,fast learning,high precision,strong generative ability,and it overcomes the shortcomings of depending on the initial value excessively and local convergence of BP Neural Network.
出处
《山西水利科技》
2011年第2期9-11,共3页
Shanxi Hydrotechnics
关键词
RBF神经网络
工业取水量
预测
RBF neural network
industrial water demand
prediction
