uncertainty

Uncertainty.

Uncertainty methods say how sure a learner is about a prediction, rather than giving a single value on its own. In data stream learning these estimates are updated as each new instance arrives, so they keep working when the underlying distribution drifts.

This module currently provides prediction intervals for regression. A prediction interval gives a range of plausible values around a point prediction.

Classes

MVE
AdaPI