Accuracy

Accuracy measures how often a model’s predictions are correct. In classification, it is calculated by dividing the number of correctly classified examples by the total number tested. The metric is easy to understand, but it can be misleading when classes are imbalanced or when different mistakes carry different costs. A fraud model that predicts every transaction as legitimate may appear highly accurate if fraud is rare while failing at its actual purpose. Accuracy should therefore be interpreted with a confusion matrix and metrics such as precision, recall, F1 score, calibration, and class-specific error rates. Evaluation data must also reflect the population and conditions where the model will operate.