FEA Add cumulative gain curve metric #18479
Conversation
| @@ -968,6 +968,75 @@ def roc_curve(y_true, y_score, *, pos_label=None, sample_weight=None, | |||
|
|
|||
| return fpr, tpr, thresholds | |||
|
|
|||
| def cumulative_gain_curve(y_true, y_score, pos_label=None): | |||
| """This function generates the points necessary to plot the Cumulative Gain for each ten percent of the samples | |||
| Note: This implementation is restricted to the binary classification task. | |||
There was a problem hiding this comment.
Each line should be less than 80 char, and the first sentence of the docstring should be a 1 line summary.
| non-thresholded measure of decisions (as returned by | ||
| decision_function on some classifiers). | ||
| pos_label (int or str, default=None): Label considered as positive and | ||
| others are considered negative |
There was a problem hiding this comment.
See the docstring formatting of other functions in this module.
|
The original issue #10003 is about both cumulative gains and lift curves. |
|
@reshamas and I fixed Doc test errors |
|
@Alit10 thanks for your work so far and for your patience! If you are still interested in working on this do you mind synchronizing with upstream? Thanks! |
5875a77 to
d538225
Compare
|
@cmarmo Hello, Done ! Is everything ok for the merge ? |
|
Thanks @Alit10 ! I believe something went wrong with the synchronization. |
Yes, it is. |
Reference Issues/PRs
Closes #10003
What does this implement/fix? Explain your changes.
Add a new metric for binary classification known as cumulative_gain_curve
Any other comments?