select_optimal: Select prediction models in an optimal fashion for the...
In maxwestphal/SEPM: Statistical Evaluation of Prediction Models

select_optimal

R Documentation

Select prediction models in an optimal fashion for the evaluation study

Description

Select prediction models in an optimal fashion for the evaluation study

Usage

select_optimal(
  comparison,
  method_select = "few",
  method_order = "param",
  target_order = "prob",
  combine_order = "min",
  max_models = Inf,
  n_eval = 100,
  prev_eval = 0.5,
  n_val = NA,
  prev_val = NA,
  rdm = TRUE,
  threshold = c(0.75, 0.75),
  batch_size = 10,
  max_iter = 50,
  target_tol = 1e-04,
  method_ext = "basic",
  method_pred = "mbeta_approx",
  return_simvals = FALSE,
  steady_plot = FALSE,
  save_plot = FALSE,
  info_plot = FALSE,
  ylim_plot = NULL,
  ...
)

Arguments

`comparison`	SEPM.comparison
`method_select`	"few" or "max"
`method_order`	"param" or "sample"
`target_order`	character, see `SIMPle::order_dist`
`combine_order`	character, see `SIMPle::order_dist`
`max_models`	integer, maximum number of models
`n_eval`	evaluation sample size
`prev_eval`	evaluation prevalence
`n_val`	validation sample size
`prev_val`	validation prevalence
`rdm`	logical, simple random sampling (TRUE) or case-control sampling (FALSE)
`threshold`	numeric, length identical to number of groups
`batch_size`	integer
`max_iter`	integer
`target_tol`	numeric, tolerance parameter
`method_ext`	"basic" or "cov"
`method_pred`	"mbeta_approx" or "rmvbin"
`return_simvals`	logical
`steady_plot`	logical
`save_plot`	logical
`info_plot`	logical
`ylim_plot`	numeric, length 2
`...`	further arguments (currently ignored)