A confusion matrix shows in a matrix the number of true positives, true negatives, false positives, and false negatives. A confusion matrix provides the information needed to evaluate models performance on a classification problem.