Prev: data acquisition using NI 9172 through matlab
Next: add zero cells to matrix
From: Charles on 19 May 2010 08:55
Hi all,

I am training a network using the 'newff' function. The output so far is good but for one issue. The y-axis of the output represent percentage (max=100) while the x-axis represent some other variable. The problem is that the output of the algorithm sometimes gets to 120% which is not possible with the type of data that I have. Is there a way to put a constraint to limit the output to 100%? Thanks.

Charles
From: Greg Heath on 20 May 2010 08:02
On May 19, 8:55 am, "Charles " <nkwos...(a)yahoo.com> wrote:
> Hi all,
>
> I am training a network using the 'newff' function. The output so far is good but for one issue. The y-axis of the output represent percentage (max=100) while the x-axis represent some other variable. The problem is that the output of the algorithm sometimes gets to 120% which is not possible with the type of data that I have. Is there a way to put a constraint to limit the output to 100%? Thanks.

y = 100*sim(net,x);

where

net = newff(p,t/100,{'tansig' 'logsig'});

Hope this helps.

Greg
 | 
Pages: 1
Prev: data acquisition using NI 9172 through matlab
Next: add zero cells to matrix