|View source on GitHub|
tf.contrib.metrics.recall_at_precision( labels, predictions, precision, weights=None, num_thresholds=200, metrics_collections=None, updates_collections=None, name=None, strict_mode=False )
recall_at_precision function creates four local variables,
tp (true positives),
fp (false positives) and
fn (false negatives)
that are used to compute the
recall at the given
precision value. The
threshold for the given
precision value is computed and used to evaluate the
For estimation of the metric over a stream of data, the function creates an
update_op operation that updates these variables and returns the
update_op increments the
fn counts with the
weight of each case found in the
None, weights default to 1. Use weights of 0 to mask values.
labels: The ground truth values, a
Tensorwhose dimensions must match
predictions. Will be cast to
predictions: A floating point
Tensorof arbitrary shape and whose values are in the range
precision: A scalar value in range
Tensorwhose rank is either 0, or the same rank as
labels, and must be broadcastable to
labels(i.e., all dimensions must be either
1, or the same as the corresponding
num_thresholds: The number of thresholds to use for matching the given
metrics_collections: An optional list of collections that
recallshould be added to.
updates_collections: An optional list of collections that
update_opshould be added to.
name: An optional variable_scope name.
strict_mode: If true and there exists a threshold where the precision is above the target precision, return the corresponding recall at the threshold. Otherwise, return 0. If false, find the threshold where the precision is closest to the target precision and return the recall at the threshold.
recall: A scalar
Tensorrepresenting the recall at the given
update_op: An operation that increments the
fnvariables appropriately and whose value matches
labelshave mismatched shapes, if
Noneand its shape doesn't match
predictions, or if
precisionis not between 0 and 1, or if either
updates_collectionsare not a list or tuple.