Delayed outputs with respect to the original NarX

Very good, I made remove delay and failure is not my concept what I'm looking for. My outputs two periods are delayed with respect to the original data. If you would tell me how to upload photos to post the chart to show you that what you saw more clearly, because I can not think of any way how. I thought that doing remove delay, the horizontal axis is the predicted data shift left me twice so get out a few steps earlier than the original data, but as I have found it does not. How could I fix it?.Would be varying the number of delays?, ie if I have ID = FD = 1:2 and 1:2 changing these parameters?

thank you very much

[Azzi Abdelmalek]

Where can I upload images and files for use on MATLAB Answers?

[Greg Heath]

Use one of MATLAB's nndatasets

help nndatasets.

Answer 1

[Greg Heath]

Please be more specific. Use one of MATLABs nndatasets and post relevant code.

[FRANCISCO]

Very good, I think the graph as shown in the following link.

https://www.rapidshare.com/files/2566312981/graficas.jpg

I summarize what I want to ask: The blue line is 51 closeloop and predictions on the red line are the actual values of the objectives in that period. The data used are outside the training sample, ie, whether for training / validation and test data use of1: 2000, for determining data used in closeloop 2000:2100 (this is an example to explain that I did not use data training sample). The problem is that the predicted outputs are delayed with respect to the original and I would get these outputs one time step earlier, but do not know how to do it because I used the Remove delay, but I have resolvido the problem. My code is mostly found in the following link:

http://www.mathworks.es/matlabcentral/answers/65925-as-used-in-closed-loop-network-narx-reality

Although I have made the appropriate changes and clarifications have changed, to give you an idea of the code that I developed while not exactly the text in the link. I'd like to specifically respondierais question I have at the beginning of the post as to what the code and I understand and I have made the appropriate changes. I'll put the code used to remove delay In case of emergency I made any error:

if true
  % code

netc = closeloop(net);

netc.name = [net.name ' - Closed Loop'];

view(netc)

NumberOfPredictions = 50;

s=cell2mat(inputSeries);

t4=cell2mat(targetSeries);

a=s(1:8,2098:2100);

b=p2(1:8,1:50);

newInputSeries=[a b];

c=t4(1,2099:2100);

d=nan(1,51);

newTargetSet=[c d];

newInputSeries=tonndata(newInputSeries,true,false);

newTargetSet=tonndata(newTargetSet,true,false);

[xc,xic,aic,tc] = preparets(netc,newInputSeries,{},newTargetSet);

yPredicted = sim(netc,xc,xic,aic);

w=cell2mat(yPredicted);

nets = removedelay(net,1);

nets.name = [net.name ' - Predict One Step Ahead'];

view(nets)

entradas=newInputSeries;

quitarretraso=[c yPredicted];

quitarretraso=tonndata(quitarretraso,true,false);

[xs,xis,ais,ts] = preparets(nets,entradas,{},objetivos);

ys = nets(xs,xis,ais);

w1=cell2mat(ys);

end

Mainly I'd like to be comentaraís to this delay in the graph and how I can fix it. The accuracies, but do not put the code of them have already been evaluated and I have no doubts.

thank you very much

[FRANCISCO]

Put another link to view the graph as I am not very familiar with rapidshare:

http://www.subirimagenes.net/i/13032212234561646.jpg

Many thanks