正交尺度小波网络及在非线性经济系统预测中的应用  

The Application of Wavelet Neural Network Based on Orthogonal Scaling Function in Nonlinear Macroeconomic Forecasting

作　　者：张新红[1] 

机构地区：[1]华侨大学经济管理学院,福建泉州362011

出　　处：《运筹与管理》2002年第6期99-103,共5页Operations Research and Management Science

基　　金：国家自然基金资助(40271054)

摘　　要：本文提出一种用正交尺度函数代替RBF网络中的激活函数的小波网络,给出相应小波网络学习算法;并以天津市国内生产总值为样本进行宏观经济模拟预测。In this paper, a waveletbased neural network is given. The structure of this network is similar to that of the radial basis function(RBF)network, except that here the radial basis functions are replaced by orthogonal scaling functions. Furthermore, the author proposes the algorithms for the wavelet network.Finally,a macroeconomic forecasting based on the sample of GDP in Tianjing is presented, which shows that the forecasting error is less than that of BP network.

关 键 词：小波神经网络 正交小波 宏观经济预测 非线性经济系统 

分 类 号：F201[经济管理—国民经济] TP183[自动化与计算机技术—控制理论与控制工程]