关于预测的MATLAB编程

clear
clc
close
t=[0:0.025:20.025];
Y=sin(t)*100+12.5+20*rand(size(t));Y=Y'
plot(Y)
figure
n=length(Y);
n1=500;
y=Y./max(Y);
(导入数据就是附件）
//get the trainning and testing data
train_data as inout for training the NN
train_target as output for training the NN
test_data as inout for testing the NN
test_target as output for testing the NN

m=20;//define a moving window e.g. 20
which is also the input number of nodes for the NN
it is also used to generate the train and the test date
load exdata
//How can you use the moving window concept to form these training the test data sets?