mimic
04-19-2005, 08:02 PM
I have a few questions concerning neural nets in general. The first would be how to do you get even a vague idea of the number of hidden nodes? Should the number of hidden nodes be greater or lesser than input nodes? The second question, well third, would be how would you feed a variable length array into a input nodes? Would simply padding the front of the array with 0's suffice? Thank you for your thoughts.