How can I get the function y(t) from NARX problem in Neural Network Time Series Tool?

How can I get the function y(t) from NARX problem in Neural Network Time Series Tool?

Answer 1

[Greg Heath]

Since your model is not nonlinear, just use NARXNET with no hidden nodes.

 X   = con2seq(giveninput);
 T   = con2seq(giventarget);
 net = narx(1:2,1:2,[]);

The rest follows from the documentation examples

 help narxnet
 doc narxet

Hope this helps

Thank you for formally accepting my answer

Greg

Answer 2

[Greg Heath]

See one of my Narx posts in the NEWSGROUP or ANSWERS:

 greg narnet

Hope this helps.

Thank you for formally accepting my answer

Greg

[Fabio Retorta]

I have tried to found the answer in your post. What I really like to know is how I can get the coeficients of y(t_1)..., u(t_1)..., of the y(t)

[Greg Heath]

Not exactly sure what you mean by "the" coefficients. The net is a nonlinear transformation. The coefficients corresponding to the inputs of the hidden node neurons are given by the input weight matrix IW = net.IW.

Hope this helps.

Thank you for formally accepting my answer

Greg

[Fabio Retorta]

In this function y(t) = a1.y(t-1)+a2.y(t-2)...+b1.u(t-1)+b2.u(t-2)

the coeficients ate a1, a2, b1 and b2. If I know this terms I can make a prediction.

Using the toolbox I don't have this fucntion. I want to know if I can get these coeficients