pystra.plotting.plot_learning_history#

plot_learning_history(result, *, band='probability', reference_probability=None, ax=None)[source]#

Plot exploratory Pf against cumulative true-model evaluations.

Parameters:

  • result (ActiveLearningResult) – Snapshot with a nonempty history. The title retains its stopping status. Historical estimates use each exploration estimator’s sampling measure; the independent final estimate is not substituted into this curve.

  • band (str | None) – Plot stored mean±2std probability sensitivity or actual bootstrap Pf ranges. Neither is a confidence interval or a bound on true Pf. Missing bootstrap ranges leave gaps; entirely absent ranges raise.

  • reference_probability (float | None) – Independently obtained probability in [0, 1], drawn for comparison.

  • ax (Axes | None) – Existing axes. No estimation or model evaluation is performed.

Returns:

figure, ax – Figure and axes, without displaying them.

Return type:

tuple[Figure, Axes]