get_pilot_covariance

pyapprox.control_variate_monte_carlo.get_pilot_covariance(nmodels, variable, model_ensemble, npilot_samples)[source]
Parameters
nmodelsinteger

The number of information sources

variablepyapprox.variable.IndependentMultivariateRandomVariable

Object defining the nvar uncertain random variables. Samples will be drawn from its joint density.

model_ensemblecallable

Function with signature

model_ensemble(samples) -> np.ndarray (nsamples,1)

where samples is a np.ndarray with shape (nvars+1,nsamples)

npilot_samplesinteger

The number of samples used to compute correlations

Returns
cov_matrixnp.ndarray (nmodels,nmodels)

The covariance between each information source

pilot_samplesnp.ndarray (nvars+1,nsamples)

The samples used to evaluate each information source when computing correlations

pilot_valuesnp.ndarray (nsamples,nmodels)

The values of each information source at the pilot samples