PRIOR_SAMPLED_INITIALIZATION#

Name#

PRIOR_SAMPLED_INITIALIZATION

Purpose#

Generates the initial parameter values from the prior, restricted to the initialization range, for sampling with the MCMC or MPFIT algorithms.

Calling Sequence#

parameter_start = prior_sampled_initialization(priors, Nmodels [, sigma_start=sigma_start])

Inputs#

priorsstructure

A structure containing the prior hyper-parameters and initialization range. (See generate_prior_struct.pro for details and contents.)

Nmodelsint, float, or double scalar

The number of model parameter sets to initialize.

Output#

parameter_startdouble array(Nparam, Nmodels)

Initial parameter values for the model(s) as sampled from each prior distribution within initialization range.

Optional Output#

sigma_startdouble array(Nparam)

Initial proposal distribution sigma values for each parameter as determined from half of the 16th and 84th percentile width of each prior distribution.

Modification History#

  • 2022/06/16: Created (Keith Doore)

  • 2022/08/02: Updated to use initialization range to initialize within prior (Keith Doore)