What is the difference between recurrent neural network and NAR and NARX networks? Is there any good literature available to understand the difference between them, especially the training algorithms in MATLAB backend implementation. Thanks

*

Answer 1

[Greg Heath]

Recurrent networks are a general category of networks that feed output or intermediate signals BACKWARD toward a PREVIOUS point in the network. That point can be the input or just another intermediate point.

NARNET and NARXNET are only two such examples.

Use the help and doc commands for explanations and examples.

e.g.

   help narxnet
   doc  narxnet

For examples search BOTH the NEWSREADER and ANSWERS

Hope this helps.

Greg

P.S. Is this significantly different from what you learned from Wikipedia?