pystra.active_learning.results.ActiveLearningResult#

Use it: User guide · Worked example · Theory

class ActiveLearningResult(estimate, converged, status, n_limit_state_evaluations, history)[source]#

Bases: object

Snapshot including unfinished estimates, diagnostics and stopping status.

estimate identifies the final estimator and its uncertainty semantics. Probability, beta, sampling diagnostics and n_estimation are exposed as convenient properties of that record. converged means the configured stopping and sampling criteria passed; it does not guarantee accuracy or discovery of all failure regions. History contains exploratory probability diagnostics, separate from the independent final estimate.

Methods

property failure_probability: float#: Final estimator’s conditional failure probability.

property beta: float#: Normal-equivalent index of the final probability.

property sampling_cov: float#: Final estimator’s conditional sampling coefficient of variation.

property sampling_interval: tuple | None#: Final estimator’s conditional sampling interval, if available.

property n_estimation: int#: Number of surrogate classifications used by the final estimator.