hi,

have a look at:

https://github.com/scikit-learn/scikit-learn/pull/1726

Cheers,
Alex

On Fri, Mar 1, 2013 at 11:31 AM, Tadej Janež
<[email protected]> wrote:
> Hi!
>
> I've been performing some tests with KFold cross validation and
> encountered a strange behavior:
>
>>>> from sklearn import cross_validation
>>>> list(cross_validation.KFold(14, 5, indices=True, shuffle=True,
> random_state=32))
> [(array([13,  2, 12,  9,  1, 10,  4,  3,  8,  6,  5, 11]), array([0,
> 7])), (array([ 0, 13, 12,  9,  1, 10,  4,  3,  8,  6,  5,  7]),
> array([ 2, 11])), (array([ 0,  2, 12,  9,  1, 10,  4,  3,  6,  5, 11,
> 7]), array([13,  8])), (array([ 0, 13,  2, 12,  1, 10,  4,  3,  8,  5,
> 11,  7]), array([9, 6])), (array([ 0, 13,  2,  9,  8,  6, 11,  7]),
> array([12,  1, 10,  4,  3,  5]))]
>
> Since I was performing a 5-fold cross validation on 14 examples, I would
> have expected the first 4 folds to have 11 training indices and 3
> testing indices and the last fold to have 12 training indices and 2
> testing indices.
>
> Does anyone have any explanation for this? Is this an expected behavior?
>
> Thanks and best regards,
> Tadej
>
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to