In sklearn.metrics.auc
documentation the auc
score is discussed, but this is different from the regular roc_auc_score
. I see no description of this, what is it and what is it used for?
sklearn.auc
is a general fuction to calculate the area under a curve using trapezoid rule. It is used to calculate sklearn.metrics.roc_auc_score
.
To calculate roc_auc_score, sklearn evaluates the false positive and true positive rates using the sklearn.metrics.roc_curve
at different threshold settings. Then it uses sklearn.metrics.auc
to calculate the area under the curves, and finally returns their average binary score.