On 5 May 2014 16:43, Gael Varoquaux <gael.varoqu...@normalesup.org> wrote:
> > Previously when discussing such things, Olivier suggested that the most
> > flexible solution is a dict per validation. A dict per validation, or a
> > dict of arrays, can also easily be converted to a Pandas DataFrame.
>
> I think that a dict of arrays is the best option, given that it maps well
> to a Pandas DataFrame, and this will make manipulation of these results
> easiest.
>
> Gaël
>
Pandas will accept an array of dicts, which may be easier to manipulate
without Pandas (e.g. it reflects the current namedtuples). I think Olivier
also suggested it could handle a change in the number of folds (as one
needs more precise validation), missing data, etc. And it corresponds to
the MongoDB records that hyperopt outputs, for example. So I'm leaning
towards array/list of dicts.
------------------------------------------------------------------------------
Is your legacy SCM system holding you back? Join Perforce May 7 to find out:
• 3 signs your SCM is hindering your productivity
• Requirements for releasing software faster
• Expert tips and advice for migrating your SCM now
http://p.sf.net/sfu/perforce
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
