Specify input for NARX

   I am following the example about NARX (autogressive nerual network with external inputs)

http://www.mathworks.com/help/nnet/ug/design-time-series-narx-feedback-neural-networks.html

What I want in my study case is that the external input need to be specified as a matrix. The manual already showed a case to address multiple external variables(which is a vector). The question is how I can let the input be a matrix?

Thanks a lot!!

Answer 1

[Greg Heath]

Inputs are assumed to be matrices containing N I-dimensional vectors corresponding to N O-dimensional target/output vectors.

 [ I N ] = size(input)
 [O N ] = size(target)

If you have an rxc dimensional matrix, A, that is supposed to be a single input

 input = A(:)
 I = r x c
 N = 1

Hope this helps.

Thank you for formally accepting my answer

Greg