sklearn.metrics.auc — scikit-learn 0.20.4 documentation (original) (raw)

x : array, shape = [n]

x coordinates. These must be either monotonic increasing or monotonic decreasing.

y : array, shape = [n]

y coordinates.

reorder : boolean, optional (default=’deprecated’)

Whether to sort x before computing. If False, assume that x must be either monotonic increasing or monotonic decreasing. If True, y is used to break ties when sorting x. Make sure that y has a monotonic relation to x when setting reorder to True.

Deprecated since version 0.20: Parameter reorder has been deprecated in version 0.20 and will be removed in 0.22. It’s introduced for roc_auc_score (not for general use) and is no longer used there. What’s more, the result from auc will be significantly influenced if x is sorted unexpectedly due to slight floating point error (See issue #9786). Future (and default) behavior is equivalent to reorder=False.