[argyllcms] Re: Randomized GCV for RSPL
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sun, 02 Jul 2006 13:36:07 +1000
Ben Goren wrote:
On 2006 Jun 24, at 7:50 AM, Gerhard Fuernkranz wrote:
> the attached patch adds the computation of randomized GCV
> (Generalized Cross Validation) to the RSPL code. Please take a
> look, whether you also find it useful.
Just to be sure...this /does/ do exactly what you were nudging me
to write in Perl a few weeks back, no?
I've had a little play with it, and it does some interesting things,
but it doesn't seem bullet proof. It tended to under smooth
a couple of the profile I tried it out on, so maybe it needs
some more investigation as to how reliable such an approach is
likely to be.
Graeme Gill.
- References:
- [argyllcms] Re: Randomized GCV for RSPL
- From: Ben Goren
Other related posts:
- » [argyllcms] Randomized GCV for RSPL
- » [argyllcms] Re: Randomized GCV for RSPL
- » [argyllcms] Re: Randomized GCV for RSPL
On 2006 Jun 24, at 7:50 AM, Gerhard Fuernkranz wrote:
> the attached patch adds the computation of randomized GCV > (Generalized Cross Validation) to the RSPL code. Please take a > look, whether you also find it useful.
Just to be sure...this /does/ do exactly what you were nudging me to write in Perl a few weeks back, no?
I've had a little play with it, and it does some interesting things, but it doesn't seem bullet proof. It tended to under smooth a couple of the profile I tried it out on, so maybe it needs some more investigation as to how reliable such an approach is likely to be.
- [argyllcms] Re: Randomized GCV for RSPL
- From: Ben Goren