Plot a confusion matrix heatmap — plot_confusion

Creates a heatmap with counts overlaid for a confusion matrix (table or matrix) of truth vs. predicted classes.

Usage

plot_confusion_matrix(
  conf_mat,
  title = "Confusion matrix",
  subtitle = NULL,
  cv_folds = NULL,
  aggregated = TRUE,
  fill_low = "#f0f9e8",
  fill_high = "#08589e",
  plot_marginals = TRUE
)

Arguments

conf_mat: A confusion matrix as `table`, `matrix`, or data frame with columns `truth`, `pred`, and `Freq`.
title: Plot title.
subtitle: Optional subtitle for the plot. If `NULL`, an automatic subtitle is generated (aggregation flag, CV folds if provided, and total `n`).
cv_folds: Optional number of cross-validation folds to display in the subtitle. Use `NULL` to omit.
aggregated: Logical. If `TRUE`, indicates the confusion matrix is summed/aggregated (used for subtitle text only).
fill_low: Low-end color for the fill gradient.
fill_high: High-end color for the fill gradient.
plot_marginals: Logical. If `TRUE`, add an extra "TOTAL" row/column showing marginal totals and the grand total.

Value

A `ggplot` object.

Examples

conf <- matrix(
  c(18, 2, 3, 17),
  nrow = 2,
  dimnames = list(truth = c("A", "B"), predicted = c("A", "B"))
)
plot_confusion_matrix(conf)