Skip to contents

Run MAGMA model

Source: R/magma_mdl.R
magmatize_mdl.Rd

Run MAGMA model

Usage

magmatize_mdl(
  dat_in,
  nreps,
  nburn,
  thin,
  nchains,
  nadapt = 50,
  keep_burn = FALSE,
  age_prior = "flat",
  cond_gsi = TRUE,
  file = NULL,
  seed = NULL,
  iden_output = FALSE
)

Arguments

dat_in

Input data list.

nreps

Amount of simulations to run.

nburn

Number of burn-in simulations to discard.

thin

At what interval to keep the simulations.

nchains

Number of independent MCMC chains run in the simulation.

nadapt

Amount of warm-up/adapt runs before the simulation (only for fully Bayesian mode).

keep_burn

Logical (default = FALSE). To keep the burn-ins in the output or not.

age_prior

Option to adjust prior weight on the age class proportions.

  • flat (default): conventional setup that puts a flat prior on the age proportions.

  • zero_out: prior weight will concentrate on the major age groups that are observed in metadata (i.e., "zero out" the unobserved age classes).

  • weak_flat: a less influential flat prior than the conventional one

cond_gsi

Logical (default = TRUE). Option to use conditional GSI model. See vignette for details.

  • TRUE: run MAGMA with a hybrid algorithm of conditional GSI and fully Bayesian.

  • FALSE: run MAGMA with fully Bayesian algorithm.

file

File path for saving the output. The default is NULL for not saving the output.

seed

Option to initialize a pseudo-random number generator (set random seed) so the output can be reproduced exactly. Just pick a seed number and make note of it for future reference. The default is NULL.

iden_output

Option to have trace history for individual group membership assignments included in the final output. Default is FALSE.

Value

A list object contains:

  • the raw output of MAGMA as a list/multi-way array that need to be further summarized using summary functions,

  • specifications for the model run (information needed for summary),

  • and individual group membership assignment history (optional).

Examples

if (FALSE) { # \dontrun{
# format data
wd <- getwd() # path to data folder
magma_data <- magmatize_data(wd = wd, save_data = FALSE)

# model run
magma_out <- magmatize_mdl(magma_data, nreps = 50, nburn = 25, thin = 1, nchains = 2)
} # }