If you are treating your Logistic Regression output as binary (i.e. not
using predict_proba or decision_function), could you please provide the
confusion matrix?
On 26 November 2015 at 05:06, Herbert Schulz wrote:
> Hi, i think i have some "missunderstanding" due to the classification
> metric in scikit-learn
>
>
>
> i have a 2 class problem it is 1.0 or 2.0
>
>
> precisionrecall f1-score support
>
> 1.0 0.86 0.76 0.81 254
> 2.0 0.49 0.65 0.5691
>
> avg / total 0.76 0.73 0.74 345
>
>
> Specificity: [ 1.* 0.35164835* 0.]
> recall,tpr,sensitivity [ 0. * 0.24015748* 1.]
>
>
> # this part is manually computed ( precision, sens, spec, ballanced
> accuracy )
>
> logistic regression 0.86,* 0.76, 0.65,* 0.7
>
>
>
> The part with:
>
> Specificity: [ 1. 0.35164835 0.]
> recall,tpr,sensitivity [ 0. 0.24015748 1.]
>
> are computed with
>
> fpr, tpr, thresholds = metrics.roc_curve(expected, predi,
> pos_label=1)
> print "Specificity:", 1-fpr
> print "recall,tpr,sensitivity",tpr
>
> Why is th speceficity for 1-fpr are computed wtih [ 1.
> 0.35164835 0.]
>
> and not 0.65 ?
>
> Same with recall
>
>
>
>
>
>
>
>
>
>
>
> --
> Go from Idea to Many App Stores Faster with Intel(R) XDK
> Give your users amazing mobile app experiences with Intel(R) XDK.
> Use one codebase in this all-in-one HTML5 development environment.
> Design, debug & build mobile apps & 2D/3D high-impact games for multiple
> OSs.
> http://pubads.g.doubleclick.net/gampad/clk?id=254741551=/4140
> ___
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
--
___
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general