EnsembleSample objects are generated using the generate_sample function.
ensemble_fitAn EnsembleFit object containing the fitted ensemble model.
mleAn array of dimension \(T \times (M + 2)\times N_{sample}\) containing MLE point estimates from the ensemble_fit object, where \(T\) is the total time, \(M\) is the number of simulators and \(N_{sample}\) is the number of samples. For each time step, the tth element of the array is a matrix where each column is a sample and the rows are the variables:
$$\left( y^{(t)}, \eta^{(t)}, z_1^{(t)}, z_2^{(t)}, \ldots, z_M^{(t)}\right)'$$
where \(y^{(t)}\) is the ensemble model's prediction of the latent truth value at time \(t\),
\(\eta^{(t)}\) is the shared short-term discrepancy at time \(t\),
\(z_i^{(t)}\) is the individual short-term discrepancy of simulator \(i\) at time \(t\).
samplesAn array of dimension \(T \times (M + 2)\times N_{sample}\) containing samples from the ensemble_fit object, where \(T\) is the total time, \(M\) is the number of simulators and \(N_{sample}\) is the number of samples. For each time step, the tth element of the array is a matrix where each column is a sample and the rows are the variables:
$$\left( y^{(t)}, \eta^{(t)}, z_1^{(t)}, z_2^{(t)}, \ldots, z_M^{(t)}\right)'$$
where \(y^{(t)}\) is the ensemble model's prediction of the latent truth value at time \(t\),
\(\eta^{(t)}\) is the shared short-term discrepancy at time \(t\),
\(z_i^{(t)}\) is the individual short-term discrepancy of simulator \(i\) at time \(t\).
EnsembleSample, generate_sample