Prepare a bootstrap tibble — make_boot

Creates bootstrap datasets and returns corresponding nm objects. Requires the necessary rsample splitting objects to be present. See examples.

make_boot_datasets(
  m,
  samples = 10,
  data_folder = file.path(nm_dir("derived_data"), "bootstrap_datasets"),
  overwrite = FALSE,
  id_var = "ID",
  ...
)

Arguments

m: An nm object.
samples: Number of samples.
data_folder: Folder (relative path) to store datasets.
overwrite: Logical (default = FALSE). Overwrites previous files.
id_var: Character (default = "ID"). Name of ID column in dataset.
...: Arguments passed to fill_input().

Value

A tibble with samples rows and an nm object object column m for execution of the bootstrap.

Examples


## The following only works inside an NMproject directory structure and
## and requires NONMEM installed
if (FALSE) {

# create example object m1 from package demo files
exdir <- system.file("extdata", "examples", "theopp", package = "NMproject")
m1 <- new_nm(run_id = "m1", 
             based_on = file.path(exdir, "Models", "ADVAN2.mod"),
             data_path = file.path(exdir, "SourceData", "THEOPP.csv"))

d <- input_data(m1)

## in your dataset production script
d <- d %>%
  mutate(
    WT_C = cut(WT, breaks = 2, labels = FALSE),
    STRATA = paste(SEX, WT_C, sep = "_")
  )

d_id <- d %>% distinct(ID, STRATA)

set.seed(123)

## create large set of resamples (to enable simulation to grow
## without ruining seed)
bootsplits <- rsample::bootstraps(d_id, 100, strata = "STRATA")

dir.create("DerivedData", showWarnings = FALSE)
bootsplits %>% saveRDS("DerivedData/bootsplit_data.csv.RData")

## In a model development script, the following, performs a
## 100 sample bootstrap of model m1

m1_boot <- m1 %>% make_boot_datasets(samples = 100, overwrite = TRUE)

m1_boot$m %>% run_nm()

## the following bootstrap template will wait for results to complete
m1_boot$m %>% nm_list_render("Scripts/basic_boot.Rmd")
}