A table showing how often a classifier correctly or incorrectly predicts each category, revealing systematic biases in predictions.