|TensorFlow 1 version||View source on GitHub|
Approximates the AUC (Area under the curve) of the ROC or PR curves.
See Migration guide for more details.
tf.keras.metrics.AUC( num_thresholds=200, curve='ROC', summation_method='interpolation', name=None, dtype=None, thresholds=None, multi_label=False, num_labels=None, label_weights=None, from_logits=False )
Used in the notebooks
|Used in the tutorials|
The AUC (Area under the curve) of the ROC (Receiver operating characteristic; default) or PR (Precision Recall) curves are quality measures of binary classifiers. Unlike the accuracy, and like cross-entropy losses, ROC-AUC and PR-AUC evaluate all the operational points of a model.
This class approximates AUCs using a Riemann sum. During the metric accumulation phrase, predictions are accumulated within predefined buckets by value. The AUC is then computed by interpolating per-bucket averages. These buckets define the evaluated operational points.
This metric creates four local variables,
false_negatives that are used to compute the AUC.
To discretize the AUC curve, a linearly spaced set of thresholds is used to
compute pairs of recall and precision values. The area under the ROC-curve is
therefore computed using the height of the recall values by the false positive
rate, while the area under the PR-curve is the computed using the height of
the precision values by the recall.
This value is ultimately returned as
auc, an idempotent operation that
computes the area under a discretized curve of precision versus recall values
(computed using the aforementioned variables). The
controls the degree of discretization with larger numbers of thresholds more
closely approximating the true AUC. The quality of the approximation may vary
dramatically depending on
thresholds parameter can be
used to manually specify thresholds which split the predictions more evenly.
For a best approximation of the real AUC,
predictions should be distributed
approximately uniformly in the range 0, 1. The
quality of the AUC approximation may be poor if this is not the case. Setting
summation_method to 'minoring' or 'majoring' can help quantify the error in
the approximation by providing lower or upper bound estimate of the AUC.
None, weights default to 1.
sample_weight of 0 to mask values.
||(Optional) Defaults to 200. The number of thresholds to use when discretizing the roc curve. Values must be > 1.|
||(Optional) Specifies the name of the curve to be computed, 'ROC' [default] or 'PR' for the Precision-Recall-curve.|
(Optional) Specifies the Riemann summation method used.
'interpolation' (default) applies mid-point summation scheme for
||(Optional) string name of the metric instance.|
||(Optional) data type of the metric result.|
(Optional) A list of floating point values to use as the
thresholds for discretizing the curve. If set, the
||boolean indicating whether multilabel data should be treated as such, wherein AUC is computed separately for each label and then averaged across labels, or (when False) if the data should be flattened into a single label before AUC computation. In the latter case, when multilabel data is passed to AUC, each label-prediction pair is treated as an individual data point. Should be set to False for multi-class data.|
(Optional) The number of labels, used when
(Optional) list, array, or tensor of non-negative weights
used to compute AUCs for multilabel data. When
boolean indicating whether the predictions (
m = tf.keras.metrics.AUC(num_thresholds=3)
m.update_state([0, 0, 1, 1], [0, 0.5, 0.3, 0.9])
# threshold values are [0 - 1e-7, 0.5, 1 + 1e-7]
# tp = [2, 1, 0], fp = [2, 0, 0], fn = [0, 1, 2], tn = [0, 2, 2]
# tp_rate = recall = [1, 0.5, 0], fp_rate = [1, 0, 0]
# auc = ((((1+0.5)/2)*(1-0)) + (((0.5+0)/2)*(0-0))) = 0.75
m.update_state([0, 0, 1, 1], [0, 0.5, 0.3, 0.9],
sample_weight=[1, 0, 0, 1])
# Reports the AUC of a model outputing a probability. model.compile(optimizer='sgd', loss=tf.keras.losses.BinaryCrossentropy(), metrics=[tf.keras.metrics.AUC()]) # Reports the AUC of a model outputing a logit. model.compile(optimizer='sgd&