pystra.active_learning.learning.LearningDecision

Use it: User guide · Worked example · Theory

class LearningDecision(index, score, threshold_satisfied)

Bases: object

Selection from the supplied pool, with a diagnostic threshold flag.

index is relative to the input predictions, not the full candidate pool. score may be infinite (for example U at zero spread), but not NaN. The stopping criterion decides whether to use threshold_satisfied.

Methods