Glossary · E-commerce ML

Conformal Prediction

Definition

Conformal prediction is a model-agnostic framework for producing calibrated prediction intervals with finite-sample coverage guarantees. Applied to demand forecasting it replaces opaque point predictions with intervals that provably contain the true demand at a specified confidence level.

Rather than relying on the parametric assumptions of classical prediction intervals, conformal prediction uses a held-out calibration set to produce a non-conformity score distribution and intervals with guaranteed marginal coverage. Split conformal, cross-conformal, and adaptive conformal inference are variants. Particularly valuable for demand forecasting where downstream decisions need honest uncertainty, not false precision.

Essays on this concept

Related concepts