pystra.active_learning.learning.ExpectedFeasibility

Use it: User guide · Worked example · Theory

class ExpectedFeasibility(*, threshold=0.001)

Bases: LearningFunction

Maximum EFF with a default threshold of 1e-3 in limit-state units.

This criterion uses a Gaussian predictive distribution. Applied to bootstrap spread, that assumption is a heuristic, not a posterior model. See Bichon et al. (2008) and learning_eff().

Methods

select

Select the maximum EFF, reporting whether it is below tolerance.

select(mean, std)

Select the maximum EFF, reporting whether it is below tolerance.

Return type:

LearningDecision