Plot a Confusion Matrix from caret's confusionMatrix Object

Creates an elegant heatmap visualization of a confusion matrix using ggplot2. The plot displays predicted vs actual class counts with color-coded cells and numeric labels for easy interpretation.

Usage

plot_confusion_matrix(
  cm,
  title = "Confusion Matrix",
  xlab = "Reference",
  ylab = "Prediction",
  low_color = "white",
  high_color = "steelblue",
  text_color = "black",
  text_size = 5
)

Arguments

cm

An object of class "confusionMatrix" created by confusionMatrix from the caret package.

title

A character string for the plot title (default: "Confusion Matrix").

xlab

Label for the x-axis representing true/reference classes (default: "Reference").

ylab

Label for the y-axis representing predicted classes (default: "Prediction").

low_color

Colour gradient start for cells with low counts (default: "white").

high_color

Colour gradient end for cells with high counts (default: "steelblue").

text_color

Colour of the numeric count labels inside cells (default: "black").

text_size

Size of the cell count labels in points (default: 5).

Value

A ggplot object that can be further customized or displayed.

Note

The factor levels are automatically set to match the original order from the confusion matrix table, ensuring consistent axis ordering.

See also

confusionMatrix for creating confusion matrices

Examples

if (FALSE) { # \dontrun{
# Example 1: Basic usage
library(caret)

# Create sample predictions and references
set.seed(123)
actual <- factor(sample(c("A", "B", "C"), 100, replace = TRUE))
predicted <- factor(sample(c("A", "B", "C"), 100, replace = TRUE))

# Generate confusion matrix
cm <- confusionMatrix(predicted, actual)

# Plot with default settings
plot_confusion_matrix(cm)

# Example 2: Customized appearance
plot_confusion_matrix(
  cm,
  title = "Model Performance: Predicted vs Actual",
  low_color = "#f7fbff",
  high_color = "#08306b",
  text_color = "white",
  text_size = 4
)
} # }