[求助]怎么做sequential mode network
 

希望有高手可以指点一下，在做一道题目要求用sequential mode模拟一条曲线。 一般用batch mode还可以，sequential mode 有点没头绪了，特别是要求用50个neuron 的时候得到的曲线我觉得不太对。
-------------------------------------------------------------------------------------------------------------------------------------

clc;
clear;
close all;

%Generating training data and desired output
x=-1:0.05:1;
y=0.8*sin(2*pi*x)+cos(pi*x);

% Test the MLP, net_output is the output of the MLP, ytest is the desired output.
xtest=(2*rand(1,100)-1); %generate randon value from -1 to 1
xtest=sort(xtest','ascend')';
ytest=0.8*sin(2*pi*xtest)+cos(pi*xtest);

%define hidden neurons
n=[1 2 3 4 5 6 7 8 9 10 50];

for j=1:length(n);
%Create network
net=newff(minmax(x),[n(j),1],{'tansig' 'purelin'},'trainlm');

%Train the MLP
net.adaptFcn = 'trains'; % Specify the sequential training
net.adaptParam.passes = 1000; % Specify the number of epoch
for i = 1 : length(x) % Convert the input and output data
NP{i} = x(i); % into cell array
NY{i} = y(i);
end
net = adapt(net,NP,NY);

net_output=sim(net,xtest);

% Plot out the test results
figure;
plot(xtest,ytest,'ro');
hold on;
plot(xtest,net_output,'b-');
hold off

end