auc_ci: AUC with a confidence interval.

In funcml: Functional Machine Learning Framework

AUC with a confidence interval.

Description

Computes AUC together with a confidence interval, using pROC::ci.auc() as the backend (DeLong by default, or bootstrap).

Usage

auc_ci(
  truth,
  prob,
  positive = NULL,
  conf_level = 0.95,
  method = c("delong", "bootstrap"),
  boot_n = 2000L,
  digits = 4L
)

Arguments

truth	Factor (or coercible to factor) of true class labels.
prob	Predicted probability of the positive class, or a probability matrix/data frame with one column per class level.
positive	Optional positive class level. Defaults to the second factor level.
conf_level	Confidence level.
method	"delong" (default, analytic) or "bootstrap".
boot_n	Number of bootstrap replicates when method = "bootstrap".
digits	Number of digits numeric columns are rounded to.

Value

A one-row data frame with auc, conf_low, conf_high, conf_level, and method.

Examples

if (requireNamespace("pROC", quietly = TRUE)) {
  set.seed(1)
  truth <- factor(rbinom(100, 1, 0.4))
  prob <- runif(100)
  auc_ci(truth, prob)
}

funcml documentation built on Aug. 22, 2026, 5:08 p.m.