Data Science and Statistics

Calibration, explained

Calibration measures whether predicted probabilities match observed frequencies.
Back to Glossary

Part of the Semantic Notion technical glossary.

Short Definition

Calibration measures whether predicted probabilities match observed frequencies.

Intuition

If a model says 70 percent confidence many times, about 70 percent of those predictions should be correct.

Technical Definition

Calibration compares forecast probabilities with empirical outcome rates, often using reliability diagrams, Brier score, or expected calibration error.

Example

A weather model is calibrated if days with 30 percent rain probability rain about 30 percent of the time.

Common Misunderstandings

Accuracy and calibration are different properties.
A model can be accurate but overconfident.

Start here

Need the broader concept map?

Return to the glossary index to move from this definition into adjacent AI, ML, mathematics, and computer science terms.