Am 23.10.2012 12:38, schrieb Mathieu Blondel:
On Tue, Oct 23, 2012 at 6:43 PM, Olivier Grisel
<[email protected] <mailto:[email protected]>> wrote:
I would rather avoid `isinstance` checks to keep our API ducktype
friendly for 3rd party integration that would not like to use our
mixins for one reason or another.
But GridSearchCV will use the generic `from_estimator` method and
users should directly use the usual metric functions already defined
in sklearn.metrics. So, I don't see the point of adding a third way to
call metrics (and all that extra code) just for the sake of duck-typing.
Can any one explain to me again the use-case we are talking about?
What do we want to find out using isinstance or duck-typing?
------------------------------------------------------------------------------
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_sfd2d_oct
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
