AI

True and false positives and negatives, accuracy and the rare-class problem, precision, recall, F1, trade-offs and thresholds, and choosing metrics by cost of errors.