摘要
将小波分析与广义回归神经网络(GRNN)相融合,构建了一种小波广义回归神经网络(WGRNN)模型。该模型应用于我国粮食总产量预测,其预测结果在精度上均优于单一的GRNN预测模型和GM(1,1)灰色预测模型,既具有神经网络非线性逼近能力和自学习能力的特性,又具有小波在时、频两域表征局部特征的功能,可为粮食产量预测的定量化和智能化提供一条新途径。
Wavelet generalized regression neural network（WGRNN） model was constructed using wavelet analysis and generalized regression neural network（GRNN）.This prediction model had better precision on predicting total food yield during 2007～2008 if compared to GRNN and grey model GM（1,1）,and it did not only have the advantages of nonlinear mapping approximation ability and convenience of calculation of neural network,but also the function of showing partial characteristics on time and frequency of wavelet analysis.It would provide a new method on quantification and intelligentialization of predicting food yield.
出处
《湖北农业科学》
北大核心
2011年第10期2135-2137,共3页
Hubei Agricultural Sciences
基金
广西科学研究与技术开发计划软科学研究项目(桂科软0997003A)
广西农业科学院基本科研业务专项(200833基)
关键词
粮食产量预测
小波分析
GM(1
1)模型
广义回归神经网络
prediction of food yield
wavelet analysis
grey model GM（1
1）
generalized regression neural network
