I was all too glad to put together a patch:
https://github.com/scikit-learn/scikit-learn/pull/3580
On 21 August 2014 01:34, Vlad Niculae wrote:
> It has confused me as well, +1.
>
> It's counterintuitive and broken, in my opinion.
>
> Vlad
>
> On Wed, Aug 20, 2014 at 2:31 PM, Gael Varoquaux
>
I have no idea about recompiling.
But I found the solution.
I remove scikit-learn installed by pip and install with
https://github.com/scikit-learn/scikit-learn.
And the problem was resolved.
Thanks everybody!
*Veck Hsiao @ PLSM Lab in Dept. of CS in NCCU*
*About me: http://fbukevin.github.io
It has confused me as well, +1.
It's counterintuitive and broken, in my opinion.
Vlad
On Wed, Aug 20, 2014 at 2:31 PM, Gael Varoquaux
wrote:
>> It's been around for so long, but it's also hard to believe that anyone
>> exploited this behaviour intentionally. Shall we be bold and just fix
>> it
yes..
try recompiling numpy then scipy and then scikit-learn
Regards,
Krishna
On Wed, Aug 20, 2014 at 5:34 PM, federico vaggi
wrote:
> Are you sure you are 32/64 bit libraries uniformly?
>
>
> On Wed, Aug 20, 2014 at 9:00 AM, Veck Hsiao wrote:
>
>> Hi ! all,
>>
>> I just installed scikit-lear
> It's been around for so long, but it's also hard to believe that anyone
> exploited this behaviour intentionally. Shall we be bold and just fix
> it with a warning?
+1 on my side. It has confused many people that I work with.
-
On 20 August 2014 21:41, Gael Varoquaux
wrote:
> On Wed, Aug 20, 2014 at 01:37:36PM +0200, federico vaggi wrote:
> > Are there any reasons at all for keeping score function in its current
> form?
>
> No. I think that it is a bug. I'd like it changed, but we need to agree
> on a way that is not to
Are you sure you are 32/64 bit libraries uniformly?
On Wed, Aug 20, 2014 at 9:00 AM, Veck Hsiao wrote:
> Hi ! all,
>
> I just installed scikit-learn and followed with
> http://scikit-learn.org/stable/tutorial/basic/tutorial.html
>
> Then I faced a problem when I tried to load dataset from sklea
On Wed, Aug 20, 2014 at 01:37:36PM +0200, federico vaggi wrote:
> Are there any reasons at all for keeping score function in its current form?
No. I think that it is a bug. I'd like it changed, but we need to agree
on a way that is not to disruptive to users (or simply decide that the
current solu
Ok - now it makes sense. As long as it uses the correct one internally
during 'fit' - that's what matters.
Are there any reasons at all for keeping score function in its current form?
On Wed, Aug 20, 2014 at 12:55 PM, Joel Nothman
wrote:
> It's actually simpler than that issue, Michael. Gri
Aah yes -- thanks for this rather importance piece of information! -- so it
is using the scoring of the underlying estimator in precedence before the
gridsearchcv's own one. Since we can't really change the internal default
scorer of ExtraTreesClassifier, except by doing something horrible like,
et
It's actually simpler than that issue, Michael. GridSearchCV (and
RandomizedSearchCV) has a score method that is unintuitive. It will
generally not use the metric passed to `scoring`. But yes, in `fit`, it has
used the correct scoring metric.
IMO, it should be changed. But it's been this way since
Hi Federico,
I recall an issue at the beginning of this year stating that internally
GridSearchCV sometimes defaulted to accuracy scoring even though a
different scorer was passed. I am not sure though if this is what you have
encountered.
There is some code in
https://github.com/scikit-learn/sci
Hi everyone,
I'm working on a classification task with ExtraTreesClassifier that deals
with somewhat imbalanced datasets, so instead of using accuracy as a
metric, I'm using MCC.
However - there's some behaviour which doesn't make perfect sense to me,
for example - after doing this:
score_func =
Hi ! all,
I just installed scikit-learn and followed with
http://scikit-learn.org/stable/tutorial/basic/tutorial.html
Then I faced a problem when I tried to load dataset from sklearn.
Python 2.7.5 (
14 matches
Mail list logo
