Reference · Glossary

Confusion matrix

A table of **true vs predicted labels** — shows false positives and false negatives, not just overall accuracy.

When to use

When the cost of wrong “yes” differs from wrong “no” (spam, fraud, medical triage demos).

When not to

As the only metric when classes are wildly imbalanced — pair with precision/recall.

Example

| | Pred 0 | Pred 1 |

|--|--------|--------|

| True 0 | TN | FP |

| True 1 | FN | TP |