On Wed, 2004-06-02 at 17:43, Vojislav Kecman wrote:
> > How does scaling of the data relate to kernel selection? Can they be
> > seen as equivalent in some way?
>
> It's not related to the type of the kernel, but it is a MUST if you want to
> have good results. Just scale the inputs between -1 and 1, or to the data
> having zero mean and variance = 1.
> Don't ask me why! It is longer than average Email letter.
> Just do it, and if curious, search the papers and books for the answer.
Just to clarify, this refers to scaling each dimension of the input, (as
opposed to scaling each vector), right?
This would 'whiten' the distribution so its approx. spherical in R^n.
The question is then how to deal with inputs that are excessivly noisy
or otherwise irrelevant.
--
Andrew (Andy) W. Schmeder
<andy \at a2hd \dot com>
http://www.a2hd.com/
|