Prev: ignore specific caught errors w/ dbstop if caught error
Next: calculating returns
From: Givemore on 2 Aug 2010 19:25
Has aaanyone found an aswer to this question? am also interested in solving the same problem
Gidz
"Velimir " <velimircirovic(a)gmail.com> wrote in message <hs6tqb$edi$1(a)fred.mathworks.com>...
> I have a trained ANN and need to use it as fitness function in GA. ANN has three input parameters and one parameter used as an output. Also, it has one hidden layer with two neurons, and one output layer with one neuraon as default. I need to optimize one of the input parameters using GA. Please, can somebody help me to write m-file needed for defining a fitness function?
> I will be very gratefull.
> Thanks,
> Velimir
From: Greg Heath on 3 Aug 2010 08:03
On Aug 2, 7:25 pm, "Givemore " <give...(a)gmail.com> wrote:
> Has aaanyone found an aswer to this question? am also interested in solving the same problem
> Gidz

PLEASE DO NOT TOP POST

Did you follow up on the Weight-elimination objective
function as I advised in

http://groups.google.com/group/comp.soft-sys.matlab/msg/3b8f67e8c0d4efce?hl=en

The basic idea is to eliminate weights
that have become too small.

Greg

 | 
Pages: 1
Prev: ignore specific caught errors w/ dbstop if caught error
Next: calculating returns