Cross validation, neutal networks, doubt

Hello everybody,

I'm seriously concerned about the matlab functions that are used to make neural networks. I need to make a neural network able to get 3 sets of vectors: a training, a validation and a testing set. The problem is: the newfit/newff function gets the set with all vectors and just uses the dividerand function on it (which is too much automatic). So, I already have the 3 sets and I need to use the newfit and the train functions using the sets I've already prepared. Any suggestions?

Thanks in advance, Guilherme

[Erfan]

Hi I have the same problem. please let me know if you find it. thanks(<mailto:[e-mail] [e-mail]>)

[Erfan]

Hi I have the same problem. please let me know if you find it. thanks(<mailto:[e-mail] [e-mail]>)

Answer 1

[Greg Heath]

You can specify the subsets using

 net.divideFcn = 'divideind'

 help divideind

 doc divideind

Hope this helps

Thank you for formally accepting my answer

Greg

P.S. You might find useful examples by searching the NEWSGROUP and ANSWERS using

 divideind

or

 greg divideind

Answer 2

[Erfan]

Hi I have the same problem. please let me know if you find it. thanks(<mailto:[e-mail] [e-mail]>)

[Greg Heath]

This is a comment, not an answer. Please delete