estimate_variance_reduction

pyapprox.control_variate_monte_carlo.estimate_variance_reduction(model_ensemble, cov, generate_samples, allocate_samples, generate_samples_and_values, get_cv_weights, get_rsquared=None, ntrials=1000.0, max_eval_concurrency=1, target_cost=None, costs=None)[source]

Numerically estimate the variance of an approximate control variate estimator and compare its value to the estimator using only the high-fidelity data.

Parameters
ntrialsinteger

The number of times to compute estimator using different randomly generated set of samples

max_eval_concurrencyinteger

The number of processors used to compute realizations of the estimators, which can be run independently and in parallel.