How to make prediction from a trained NARX neural network?
3 views (last 30 days)
Show older comments
I have got the following code from a research paper which implements a NARX Neural network which trains the network using one exogenous input:
% Anp – The input time series.
% Adtds – The feedback time series.
X = tonndata(Anp,true,false);
T = tonndata(Adtds,true,false);
% 'trainlm' training function is chosen
trainFcn = 'trainlm'; % Levenberg-Marquardt backpropagation.
% Model creation
inputDelays = 1:2;
feedbackDelays = 1:2;
hiddenLayerSize = 10;
net = narxnet(inputDelays,feedbackDelays,hiddenLayerSize,'open',trainFcn);
% Training and simulation data preparation
[x,xi,ai,t] = preparets(net,X,{},T);
% Divide the data for training, validation and testing
net.divideParam.trainRatio = 70/100;
net.divideParam.valRatio = 0/100;
net.divideParam.testRatio = 30/100;
net.divideFcn = 'divideblock';
% Network training
[net,tr] = train(net,x,t,xi,ai);
% Network testing
y = net(x,xi,ai);
e = gsubtract(t,y);
performance = perform(net,t,y)
% Network view
view(net)
% Plots
figure, plotperform(tr)
figure, plottrainstate(tr)
figure, ploterrhist(e)
figure, plotregression(t,y)
figure, plotresponse(t,y)
figure, ploterrcorr(e)
figure, plotinerrcorr(x,e)
% Closed Loop Network
netc = closeloop(net);
netc.name = [net.name ' - Closed Loop'];
view(netc)
[xc,xic,aic,tc] = preparets(netc,X,{},T);
yc = netc(xc,xic,aic);
closedLoopPerformance = perform(net,tc,yc)
% Step-Ahead Prediction Network
nets = removedelay(net);
nets.name = [net.name ' - Predict One Step Ahead'];
view(nets)
[xs,xis,ais,ts] = preparets(nets,X,{},T);
ys = nets(xs,xis,ais);
stepAheadPerformance = perform(nets,ts,ys)
I am able to understand that it is training the network. But i am not able to understand how to predict output for new input data which the network has never seen before. I tried using net(input_Series) but it gives me the error that inputs are not sufficient. Could anyone please help me out?
0 Comments
Answers (1)
Greg Heath
on 17 Aug 2020
You forgot to include the intial conditions:
yz = nets(xz,xiz,aiz);
Thank you for formally accepting my answer
Greg
2 Comments
georg enyew
on 5 Feb 2021
Edited: georg enyew
on 5 Feb 2021
this problem happen to the same to me? how it could be? any one who help us i appreciated.
See Also
Categories
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!