English
资源说明：新手求助RBF神经网络数据预测问题-数据.xls
本程序用前六个数据预测下一数据，前200组数据用于训练，用后80组数据进行预测，可预测结果为一个值，请大家指教：
源程序为：
clc
clear
close all
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%每五个数据整合,
x=xlsread;
n=length
m=5;
sum=0;
j=1;
x1=zeros);
for k=1:m:n
for i=k:
sum=sum x;
end
x1=sum;
j=j 1;
sum=0;
end
j
t=1:;
figure
plot;%,'d-m'
hold on;
grid on
title;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%归一化处理
[x2,mint,maxt] = premnmx
figure
plot;%,'d-m'
hold on;
grid on
title;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%输入数据
for k=1:1:200
p_train=[x2 x2 x2 x2 x2 x2]; 5*6
t_train=x2; 5*1
end
p_train=p_train'; %6*195
t_train=t_train'; %1*195
for z=1:1:81
p_test=[x2 x2 x2 x2 x2 x2];
t_test=x2;
end
p_test=p_test';
t_test=t_test';
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%建立 RBF 网络
goal = 0.0001; % 训练误差的平方和
spread = 0.01; % 此值越大,需要的神经元就越少
MN = size; % 最大神经元数
DF = 1; % 显示间隔
net = newrb;
data_out=sim
p_mse=mse
figure
plot
hold on
plot
title;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%预测并作图
y=sim
y_mse=mse
figure
plot;
hold on;
plot
title;
legend;
本源码包内暂不包含可直接显示的源代码文件，请下载源码包。
