Drexel dragonThe Math ForumDonate to the Math Forum



Search All of the Math Forum:

Views expressed in these public forums are not endorsed by Drexel University or The Math Forum.


Math Forum » Discussions » Software » comp.soft-sys.matlab

Topic: validation in neural network
Replies: 1   Last Post: Mar 10, 2013 11:36 AM

Advanced Search

Back to Topic List Back to Topic List Jump to Tree View Jump to Tree View  
Greg Heath

Posts: 5,944
Registered: 12/7/04
Re: validation in neural network
Posted: Mar 10, 2013 11:36 AM
  Click to see the message monospaced in plain text Plain Text   Click to reply to this topic Reply

"Anurag Bishnoi" wrote in message <khh89h$n7c$1@newscl01ah.mathworks.com>...
> what is the difference between validation and testing in neural network training?
>
> what is the meaning of line " Best validation performance is 11.23334 at 10 epochs?
>
> plz explain in detail


Training has stopped because the performance on the nontraining validation set has
failed to decrease for

tr.trainParam.max_fail = net.trainParam.max_fail epochs (default = 6).

also see tr.stop.

Assuming the default, training stopped at tr.num_epochs = 16. However, the nontraining validation set is a representative of the parent population of data. Therefore, to achieve a net that performs well on the parent population (good generalization), the previous design at

tr.best_epoch = tr.num_epochs - tr.trainParam.max_fail

is chosen.

Consequently, to obtain a completely unbiased estimate of generalization performance, use

MSEtest = tr.best_tperf

instead of

MSEtest = tr.tperf(end).

Hope this helps.

Greg



Point your RSS reader here for a feed of the latest messages in this topic.

[Privacy Policy] [Terms of Use]

© Drexel University 1994-2014. All Rights Reserved.
The Math Forum is a research and educational enterprise of the Drexel University School of Education.