Problem with network NarX need help for final step

Good, I write again to see if anyone can help me as I left the last steps to fully utilize the network NarX. I expose the example I put in previous post: I've trained a network with exogenous variables NarX, I checked the accuracy, and I've done compared closeloop and now I want accuracy in predicting real-time, ie if I introduce today inputs (14/03/2012) and I want to get predictions in the next 10 days as you would? Should use the predicted outputs to make that prediction but how?. If there is some code to understand me would be helpful, or if there is any book, web .... where I could learn to perform this step. The accuracy, code, etc ... are published in previous post, so I'd like to know this step.

Thank you very much

[Greg Heath]

Don't make someone who is trying to help search for a previous post in a different thread.

Alternatives:

1. Provide all info needed in the current thread.

    a. Directly

    b. Via http link

2. Add the current post to the previous thread.

Thanks.

Greg

[FRANCISCO]

Sorry for the dismissal; attached the links where you can check my code as much relevant detail and also achieved my question still has not been resolved and I do not know how to solve it, the last step in the realization of the network NarX:

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

http://www.mathworks.es/matlabcentral/answers/30885-preparets-command-in-matlab-ver-7-9-0-r2009b

If anyone knows how to solve my doubt putting some example code so that it can understand, or know where I can find information to perform this step would greatly appreciate it.

Thank you very much

[FRANCISCO]

Currently I have matlab 2012 that I had forgotten comment.

Thank you very much

Answer 1

[Greg Heath]

If you look at the significant lags of the output autocorrelation function and the output/input crosscorrelation function, you can deduce how far ahead you might be able to predict AND the number of delays that you might have to use to accomplish that goal. In general,

y( : , t : t + d ) = f( x( : , t - id : t) , y( : , t - ld : t - 1))

For a special type of NARXNET scalar output prediction

y( t + d ) = f( x( : , t - id : t) , y( t - ld : t - 1))

for d = 10, id >= 0, ld >= 1

Whereas for a special type of TIMEDELAYNET prediction

y( t + d ) = f( x( : , t - id : t) )

PS: timedelaynet(ID) may be internally different than narxnet(ID,[])!

Hope this helps.

Thank you for formally accepting my answer

Greg

[FRANCISCO]

good, in the autocorrelation and cross correlation tells me I have 10 lag significant points ahead, and the highest peaks to arise in two lag. So how should change and what should be modified to achieve 10 times predict ahead?. Should I create another network according to narxnet where shoving it exits me back other outlets predicted to achieve 10 days predicted ahead?. Do not quite understand what I do now. thank you very much.

Answer 2

[Greg Heath]

If you want to predict exactly 10 days ahead, if possible, design a net that will do that.

Iterating a net that will predict one or two days ahead doesn't seem to be the way to go (if you can avoid it).

Better yet, design a net with a 10-dimensional output that predicts ahead for a 10-day window.

Hope this helps.

Thank you for formally accepting my answer

Greg

[FRANCISCO]

Do not quite understand what I want to explain. How should I do that? taking inputs from a current day and wanting to know the values of the 10 days?

Many thanks