get_estimator

pyapprox.multifidelity.get_estimator(estimator_types, stat, costs, max_nmodels=None, **est_kwargs)[source]
Parameters:
estimator_typeslist [str] or str

If str (or len(estimators_types==1), then return the estimator named estimator_type (or estimator_types[0])

stat_typestr

The type of statistics to compute

costsnp.ndarray (nmodels)

The computational cost of evaluating each model

stat_argslist or tuple

The arguments that are needed to compute the statistic

max_nmodelsinteger

If None, compute the estimator using all the models. If not None, find the model subset that uses at most max_nmodels that minimizes the estimator covariance.

est_kwargsdict

Keyword arguments that will be passed when creating each estimator.