NARx Neural Network inputs

Asked by Joseph on 1 Feb 2012

Latest activity Answered by Greg Heath on 23 Mar 2016

3 views (last 30 days)

I have a Narx network with 6 external inputs. These have a 1:2 delay on them. I also have a feedback loop for multi-step ahead prediction with a 1:24 delay on them. Does this mean I have 12 external inputs (x1(t),x1(t-1)...x6(t-1)) and 24 'feedback' neurons (y(t-1),y(t-2)...y(t-24)) totaling 36 input neurons?

delay=24;
inputDelays = 1:2;
feedbackDelays = 1:delay;
net = narxnet(inputDelays,feedbackDelays,hiddenLayerSize);

Thanks

0 Comments

1 Answer

Link

Answer by Greg Heath on 23 Mar 2016

Accepted Answer

Yes.

However, input nodes are not neurons.

Hidden and output nodes are neurons often associated with a nonlinear transfer function like tansig (i.e., tanh) or logsig.

Hope this helps.

Greg

You are now following this question

Joseph (view profile)