Classification & Model Evaluation Calculators

A classifier or diagnostic test makes a yes-or-no prediction, and every prediction is either right or wrong in one of four ways: true positive, false positive, true negative, or false negative. The calculators here start from those four counts and derive the measures used to judge performance.

No single measure tells the whole story. Accuracy can look high on unbalanced data even when a model misses most of the rare cases, so it is worth reading several measures together.