Parameter sweeps and model iteration#
Parameter sweeps for model calibration#
(more info) For more information on model calibration, see Calibrate simulations.
Parameter sweeps and stochasticity#
With a stochastic model (such as EMOD), it is especially important to utilize parameter sweeps, not only for calibration to data or parameter selection, but to fully explore the stochasticity in output. Single model runs may appear to provide good fits to data, but variation will arise and multiple runs are necessary to determine the appropriate range of parameter values necessary to achieve desired outcomes. Multiple iterations of a single set of parameter values should be run to determine trends in simulation output: a single simulation output could provide results that are due to random chance.