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.
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)
#> truth Class1 Class2 predicted
#> 1 Class2 0.003589243 0.9964107574 Class2
#> 2 Class1 0.678621054 0.3213789460 Class1
#> 3 Class2 0.110893522 0.8891064779 Class2
#> 4 Class1 0.735161703 0.2648382969 Class1
#> 5 Class2 0.016239960 0.9837600397 Class2
#> 6 Class1 0.999275071 0.0007249286 Class1
roc_auc(two_class_example, truth, Class1)
#> # A tibble: 1 × 3
#> .metric .estimator .estimate
#> <chr> <chr> <dbl>
#> 1 roc_auc binary 0.939