Skip to content
Back to Glossary
E

Expected Calibration Error

Model EvaluationAI Term

Definition

A metric that quantifies the average difference between an AI model's predicted probabilities and the true, observed frequencies over different confidence intervals.

In short: A metric that quantifies the average difference between an AI model's predicted probabilities and the true, observed frequencies over different confidence intervals.

Clear your comparison?

Remove all selected tools and start again.