pystra.active_learning.subset.SubsetRun

Use it: User guide · Worked example · Theory

class SubsetRun(failure_probability, probability_band, sampling_cov, n_samples, levels, converged, status)

Bases: object

One independent replication, including incomplete-run diagnostics.

sampling_cov is a within-chain approximation with an IID floor; it omits cross-level dependence. The aggregate estimator also uses variation between independent complete runs. Probability bands are computed using the same nested measure, not pooled conditional sample proportions.

Methods