pystra.active_learning.stopping.StoppingCriterion

Use it: User guide · Worked example · Theory

class StoppingCriterion(*, target_cov=0.1)

Bases: ABC

Separate surrogate convergence from final sampling precision.

Policies must not keep mutable per-run counters: use the supplied history so repeated analyses start fresh. target_cov defaults to 0.1. The default final check rejects zero/all failures and nonfinite CoV, even if the surrogate policy is satisfied. Custom policies may override either decision and must document the meaning of convergence.

Methods

accepts_estimate	Require an interior probability and adequate finite sampling CoV.
should_stop	Decide surrogate convergence from fits up to and including this fit.

abstractmethod should_stop(history)

Decide surrogate convergence from fits up to and including this fit.

Return type:

bool

accepts_estimate(estimate)

Require an interior probability and adequate finite sampling CoV.

Return type:

bool