[pymvpa] Tip of the day: MVPA_SEED

Per B. Sederberg persed at princeton.EDU
Fri Jun 20 12:17:15 UTC 2008


Ahhhh, it all makes sense in the right context...  

I was starting to worry that there were massive differences in the
classifiers due to different random seeds when the data were held
constant.  That would have been nothing to smile about :)

Best,
P

At Fri, 20 Jun 2008 13:44:43 +0200,
Michael Hanke wrote:
> 
> On Fri, Jun 20, 2008 at 07:19:17AM -0400, Per B. Sederberg wrote:
> > Yes, but why does it give rise to drastically different results with
> > smiley faces?  :)
> ;-)
> 
> The particular unit tests assumes a certain classifier performance on a
> random dataset (questionable, I know) -- and sometimes (by chance) that
> dataset is a little odd. ;-)
> 
> 
> Michael
> 
> -- 
> GPG key:  1024D/3144BE0F Michael Hanke
> http://apsy.gse.uni-magdeburg.de/hanke
> ICQ: 48230050
> 
> _______________________________________________
> Pkg-ExpPsy-PyMVPA mailing list
> Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
> http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa



More information about the Pkg-ExpPsy-PyMVPA mailing list