Thanks Andy,

I was confused because the documentation for make_scorer() doesn't show it
taking a pos_label parameter. The documentation for f1_score(), though,
does show it taking a pos_label parameter.

On Mon, May 18, 2015 at 11:32 AM, Andreas Mueller <t3k...@gmail.com> wrote:

>  You need to do:
> min_f_scorer = make_scorer(f1_score, pos_label=1)
>
> See the example here:
>
> http://scikit-learn.org/dev/modules/generated/sklearn.metrics.make_scorer.html#sklearn.metrics.make_scorer
> and
>
> http://scikit-learn.org/dev/modules/model_evaluation.html#defining-your-scoring-strategy-from-metric-functions
>
> Hth,
> Andy
>
>
>
> On 05/17/2015 11:54 AM, Adam Goodkind wrote:
>
> Hi Andy,
>
>  I'm following up on this from a while ago, because I finally got around
> to trying to implement it.
>
>  To get the minority f1 score from my grid search, I have:
> min_f_scorer = make_scorer(f1_score(pos_label=1))
>  grid_search = GridSearchCV(pipeline, parameters, verbose=1, cv=3,
> scoring=min_f1_scorer)
>
>  But the f1_score() constructor needs the y_true and y_pred arrays,
> right? I'm not sure what I'm doing wrong, since it would only have y_pred
> after the classifier was run.
>
>  Thanks again,
> Adam
>
>   On Fri, May 8, 2015 at 12:39 PM, Andreas Mueller <t3k...@gmail.com>
> wrote:
>
>>  Hi Adam.
>> You can get the f_score by setting ``GridSearchCV(...,
>> scoring='f1_score')``.
>> By default, the "positive" class if the one labeled "1":
>>
>> http://scikit-learn.org/dev/modules/generated/sklearn.metrics.f1_score.html#sklearn.metrics.f1_score
>>
>> If you want more control, you can do
>> my_f_scoring = make_scorer(f1_score, pos_label="minority_label",
>> other_arguments=that_you_want)
>> GridSearchCV(..., scoring=my_f_scoring)
>>
>> You can not compute accuracy and f1 score at the same time, though, which
>> is a known limitation, which we will fix soon.
>>
>> Cheers,
>> Andy
>>
>>
>>
>> On 05/08/2015 11:57 AM, Adam Goodkind wrote:
>>
>>  In performing a grid search, I know we can get overall accuracy, but
>> I'm also interested in the minority F score. Is there a way to get that for
>> each parameter setting combination?
>>
>>  Thanks,
>> Adam
>>
>>  --
>>  *Adam Goodkind *
>> adamgoodkind.com <http://www.adamgoodkind.com>
>> @adamgreatkind <https://twitter.com/#%21/adamgreatkind>
>>
>>
>>  
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM 
>> Insight.http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>
>>
>>
>> _______________________________________________
>> Scikit-learn-general mailing 
>> listScikit-learn-general@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> Scikit-learn-general mailing list
>> Scikit-learn-general@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>
>
>  --
>  *Adam Goodkind *
> adamgoodkind.com <http://www.adamgoodkind.com>
> @adamgreatkind <https://twitter.com/#%21/adamgreatkind>
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM 
> Insight.http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>
>
>
> _______________________________________________
> Scikit-learn-general mailing 
> listScikit-learn-general@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>


-- 
*Adam Goodkind *
adamgoodkind.com <http://www.adamgoodkind.com>
@adamgreatkind <https://twitter.com/#!/adamgreatkind>
------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to