Search All of the Math Forum:

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

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: Neural Network Training Problem
Replies: 5   Last Post: Oct 13, 2013 5:10 PM

 Messages: [ Previous | Next ]
 Greg Heath Posts: 6,387 Registered: 12/7/04
Re: Neural Network Training Problem
Posted: Oct 11, 2013 8:12 PM

"raha" wrote in message <l36qpi\$m15\$1@newscl01ah.mathworks.com>...
> "Greg Heath" <heath@alumni.brown.edu> wrote in message <l36o8p\$95q\$1@newscl01ah.mathworks.com>...
> > "raha" wrote in message <l34f4q\$f46\$1@newscl01ah.mathworks.com>...
> > > "nahal geshnizjani" <nahal.geshnizjani@gmail.com> wrote in message <l3426e\$5f6\$1@newscl01ah.mathworks.com>...
> > > > Hello,
> > > >
> > > > I was wondering if we can modify error term in nntoolbox, so instead of T-Y (T being the target and Y the predicted value) it uses F(T)-F(Y).
> > > > I appreciate your help.

> >
> > Not without revising source code.
> >
> > I don't see the point.
> >
> > Stick with F(T-Y).
> >
> > Greg

> Thanks for your reply. You see the problem is that it does not matter if T and Y are not close to each other as long as F(T) and F(Y) are.

In that case just transform the target initially.

Tnew = F(Told)

Hope this helps.

Greg

Date Subject Author
10/9/13 nahal geshnizjani
10/9/13 nahal geshnizjani
10/10/13 Greg Heath
10/10/13 nahal geshnizjani
10/11/13 Greg Heath
10/13/13 chaudhry