TP and TN indicate when the classifier is correct, whereas FP and FN indicate when the classifier is incorrect.

Error rate = FP + FN / TP + TN + FP + FN

Used for: Medical data, fraud detection

Significant majority of the negative class and minority of the
positive class

Classifier Accuracy = TP + TN / TP + TN + FP + FN

Precision = TP/(TP + FP)
= TP/P'

True Positive recognition rate: sensitivity = TP/P

F = 2 x precision x recall / precision + recall

Fß = (1 + ß^2) x precision x recall / ß^2 x precision + recall

True Negative recognition rate: sensitivity = TN/N

Recall TP/(TP+FN) = TP/P