prob-metrics: Class probability metrics

In yardstick: Tidy Characterizations of Model Performance

Class probability metrics

Description

Class probability metrics evaluate soft classification predictions where truth is a factor and estimate consists of class probability columns. These metrics assess how well predicted probabilities match the true class membership.

Input requirements

• truth: factor

• estimate / ...: numeric columns containing class probabilities

Available metrics

average_precision()

Direction: maximize. Range: [0, 1]

brier_class()

Direction: minimize. Range: [0, 1]

classification_cost()

Direction: minimize. Range: [0, Inf]

gain_capture()

Direction: maximize. Range: [0, 1]

mn_log_loss()

Direction: minimize. Range: [0, Inf]

pr_auc()

Direction: maximize. Range: [0, 1]

roc_auc()

Direction: maximize. Range: [0, 1]

roc_aunp()

Direction: maximize. Range: [0, 1]

roc_aunu()

Direction: maximize. Range: [0, 1]

See Also

class-metrics for hard classification metrics

ordered-prob-metrics for ordered probability metrics

vignette("metric-types") for an overview of all metric types

Examples

data("two_class_example")

head(two_class_example)

roc_auc(two_class_example, truth, Class1)

yardstick documentation built on April 8, 2026, 1:06 a.m.