|
|
|||||||||
|
|||||||||
| |||||||||
|
|
|
| |||||||||
![]() |
|
|
«
Previous Thread
|
Next Thread
»
|
Thread Tools | Search this Thread | Rate Thread | Display Modes |
|
#1
|
|||
|
|||
|
Neural Networking
I'm not a proper programmer, i just muck about with games etc.
At the mo, I'm trying to play with neural networks for learning, but i dont know how to hand an input that isnt boolean, ie.. 0-600 as one input. is there a way to represent these inputs as one input each rather than 600 separate ones? I hope that makes sense, thx kev |
|
#2
|
||||
|
||||
|
yes. The easiest is just to ensure that your training data is split along the same variables, ie anything that falls into the 0-600 range is a yes, anything that doesn't is a no. then you can build as many inputs as you want to have ranges for. That's the easiest. However you can do one with weighted inputs, ie free variables that have been normalized. Depends how complex you want to go.
|
![]() |
| Viewing: Dev Shed Forums > Programming Languages - More > Software Design > Neural Networking |
| Thread Tools | Search this Thread |
| Display Modes | Rate This Thread |
|
|
|
|