[pymvpa] custom partitioner

Yaroslav Halchenko debian at onerussian.com
Wed May 18 17:59:32 UTC 2016

On Wed, 18 May 2016, Wolfgang Pauli wrote:

>    Hi!
>    I found it on the mailing
>    list:A http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/2013q3/002550.html

thanks!  I guess there is should be alright, since
[d for d in Splitter('partitions').generate(pset)]
generally (if only 2 partitions) would generate two splits, first one will be the 
training, 2nd testing.

>    On Wed, May 18, 2016 at 6:41 AM, Yaroslav Halchenko
>    <debian at onerussian.com> wrote:

>      On Wed, 18 May 2016, Wolfgang Pauli wrote:

>      >A  A  Thank you for the explanation. This makes sense now. It was
>      working well
>      >A  A  within Cross-validation, but when I was calling Splitter later
>      for a
>      >A  A  manual sensitivity analysis, for which I had found a code
>      snippet, the
>      >A  A  analysis was using partitions 0 1 instead. So I got very
>      different results
>      >A  A  during cross-validation, in comparison to the later sensitivity
>      >A  A  analysis.

>      yeah -- that is worrisome... I wonder now why decided with 1, 2 and not
>      0, 1 ... heh heh... I wonder if we should change that ...

>      But which snippet you have found?A  is that something within
>      PyMVPA?
Yaroslav O. Halchenko
Center for Open Neuroscience     http://centerforopenneuroscience.org
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik        

More information about the Pkg-ExpPsy-PyMVPA mailing list