pystra.active_learning.surrogates

Use it: User guide · Worked example · Theory

Optional active-learning reliability in independent normal coordinates.

AK-MCS: Echard, Gayton and Lemaire (2011), doi:10.1016/j.strusafe.2011.01.002. EFF: Bichon et al. (2008), doi:10.2514/1.34321. Adaptive sparse bootstrap PCE: Marelli and Sudret (2018), doi:10.1016/j.strusafe.2018.06.003, using UQLab 2.2.0 selection rules. See THIRD_PARTY_NOTICES and docs/uqlab-pce-provenance.md.

Classes

EnsembleSurrogate	A surrogate that also exposes row-wise replicate predictions.
KrigingSurrogate	Matérn 5/2 Gaussian process with normalized response.
PCECandidate	A candidate degree/truncation and its final OLS error diagnostics.
PCEFitResult	Immutable selected Hermite expansion and evaluated candidate scores.
PCESurrogate	Adaptive sparse Hermite PCE with bootstrap local prediction spread.
Surrogate	Predict a scalar limit state in independent standard normal space.