RBF网络和BP网络在海水盐度建模中的比较研究

摘要

A new RBF neural network is introduced and at the same time, its' structure, feature and principium are also expatiated.Contrasting with BP neural network model, it has faster convergence and better precision when it is used in the salinity modeling.A BP neural network model is set up and trained in this paper, in order to approach compensate the effects of improve non-linearity.Test proves it is practical and dependable in the field of salinity modeling and has nice applied prospect.%介绍了RBF神经网络模型结构、特点及原理,并针对海水盐度参数具有受诸多因素影响的复杂的非线性输入输出特性,训练并建立了海水盐度的RBF(Radial Basis Function)神经网络模型,为海水盐度的预测提供了一种新的方法.与BP神经网络模型相比.该模型具有收敛速度快,精度高的优点.比较结果表明,该方法在海水盐度建模等复杂系统方面具有实用性和可靠性,并有很好的应用前景.