↧
Answer by Neil Slater for Learning an arbitrary function using a feedforward net
You need to scale the input. Neural networks work best with a limited input domain, and train badly when it is exceeded. For statistical data, you would typically scale your input to have mean 0,...
View Article
Image may be NSFW.
Clik here to view.
Clik here to view.
Learning an arbitrary function using a feedforward net
I would like to get a simple example running in matlab that will use a neural net to learn an arbitrary function from input output data (basically model identification) and then be able to approximate...
View Article