Skip to contents

Fit TMB model using nlminb

Source: R/fit_tmb.R
fit_tmb.Rd

Runs optimization on the TMB model using stats::nlminb. If specified, takes additional Newton steps and calculates standard deviations. Internal function called by fit_wham.

Usage

fit_tmb(
  model,
  n.newton = 3,
  do.sdrep = TRUE,
  do.check = FALSE,
  save.sdrep = FALSE,
  use.optim = FALSE,
  opt.control = NULL
)

Arguments

model

Output from TMB::MakeADFun.

n.newton

Integer, number of additional Newton steps after optimization. Default = 3.

do.sdrep

T/F, calculate standard deviations of model parameters? See TMB::sdreport. Default = TRUE.

do.check

T/F, check if model parameters are identifiable? Runs internal check_estimability, originally provided by the TMBhelper package. Default = TRUE.

save.sdrep

T/F, save the full TMB::sdreport object? If FALSE, only save summary.sdreport) to reduce model object file size. Default = FALSE.

use.optim

T/F, use stats::optim instead of stats::nlminb? Default = FALSE.

opt.control

list of control parameters to pass to optimizer. For nlminb default = list(iter.max = 1000, eval.max = 1000). For optim default = list(maxit=1000).

Value

model, appends the following:

model$opt

Output from stats::nlminb

model$date

System date

model$dir

Current working directory

model$rep

model$report(model$env$last.par.best)

model$TMB_version

Version of TMB installed

model$parList

List of parameters, model$env$parList(x = model$opt$par, par = model$env$last.par.best)

model$final_gradient

Final gradient, model$gr(model$opt$par)

model$sdrep

Estimated standard deviations for model parameters, TMB::sdreport or summary.sdreport)

See also

fit_wham, retro, check_estimability