fit_sraplus: Run sraplus
In DanOvando/sraplus: Run sraplus Assessments

fit_sraplus

R Documentation

Run sraplus

Description

The main function for taking an object created by format_driors and fitting a surplus production model using sraplus

Usage

fit_sraplus(
  driors,
  include_fit = TRUE,
  seed = 42,
  plim = 0.05,
  model_name = "sraplus_tmb",
  randos = "log_proc_errors",
  draws = 1e+05,
  n_keep = 2000,
  engine = "sir",
  cores = 4,
  chains = 1,
  cleanup = FALSE,
  max_treedepth = 10,
  adapt_delta = 0.8,
  estimate_shape = FALSE,
  estimate_qslope = FALSE,
  estimate_q = TRUE,
  estimate_proc_error = TRUE,
  estimate_initial_state = TRUE,
  estimate_k = TRUE,
  estimate_f = FALSE,
  learn_rate = 0.001,
  analytical_q = FALSE,
  use_baranov = TRUE,
  include_m = FALSE,
  ci = 0.89,
  try_again = FALSE,
  eps = 1e-06,
  max_time = Inf,
  eval.max = 200,
  iter.max = 150,
  rel.tol = 1e-10,
  loopnum = 2,
  newtonsteps = 0,
  tune_prior_predictive = TRUE,
  index_fit_tuner = "sir",
  refresh = 250,
  log_bias_correct = TRUE,
  workers = 1,
  thin_draws = FALSE,
  thin_rate = 0.5,
  catch_cv = 0.01,
  non_centered = FALSE,
  ...
)

Arguments

`driors`	a list of driors passed from sraplus::format_driors
`include_fit`	logical indicating whether to return the fitted object
`seed`	seed for model runs
`plim`	cutoff (in units of B/K) for hockey stick PT function
`model_name`	the name of the sraplus TMB version to be run, defaults to "sraplus_tmb"
`randos`	random effects when passing to TMB
`draws`	the number of SIR samples to run
`n_keep`	the number of SIR samples to keep
`engine`	one of 'sir','stan', or 'tmb'
`cores`	number of cores for stan fits
`chains`	number of chains for stan fits
`cleanup`	logical indicating whether to remove the compiled TMB model after running
`max_treedepth`	max_treedepth for models fit using stan
`adapt_delta`	adap_delta for models fit using stan
`estimate_shape`	logical indicating whether to estimate the shape parameter of the Pella-Tomlinson model. If FALSE shape parameter is held at the initial value, either default or supplied
`estimate_qslope`	logical indicating whether to estimate a slope parameter for q in CPUE fitting. If FALSE q_slope is held at the initial value, either default or supplied. if TRUE, `estimate_proc_error` should be set to FALSE
`estimate_q`	estimate catchability coefficient or leave fixed
`estimate_proc_error`	logical indicating whether to estimate process errors. If FALSE process errors are not included in the model
`estimate_initial_state`
`estimate_k`	estimate carrying capacity
`estimate_f`	estimate fishing mortality
`learn_rate`	learn_rate for HMC
`analytical_q`	use analytical q instead of empirical
`use_baranov`	use baranov catch equation
`include_m`	include natural mortality (deprecated)
`ci`	confidence/credible interval range for summaries
`try_again`	try fit again with adjusted starting guess
`eps`	eps for nlminb
`max_time`	max_time for nlminb
`eval.max`	eval.max for nlminb
`iter.max`	iter.max for nlminb
`rel.tol`	for nlminb
`loopnum`	for nlminb
`newtonsteps`	for TMBhelper
`tune_prior_predictive`	tune prior predictive to resolve borel's paradox
`index_fit_tuner`	"sir" tunes priors for index fit
`refresh`	refresh rate for Stan
`log_bias_correct`	correct for log-transformation bias
`workers`	number of workers for parallel processing of stan outputs
`thin_draws`	true or false thing draws
`thin_rate`	thinning rate
`catch_cv`	assumed CV for catch data when estimating catches rather than treating as data
`non_centered`	use centered or non-centered form for process errors
`...`	additional parameters