Plot bootstrap parameter distributions — plot

Creates a faceted density + histogram plot showing the distribution of parameter estimates across bootstrap samples. Optionally overlays the original (fitted) parameter estimate and confidence interval lines.

Usage

plot_bootstrap(
  bootstrap_results,
  original_estimates = NULL,
  ci = 0.95,
  parameters = NULL
)

Arguments

bootstrap_results: Bootstrap results object returned by call_pharmpy_tool(tool = "bootstrap"). Must expose a parameter_estimates attribute (a pandas DataFrame that reticulate auto-converts to an R data.frame).
original_estimates: Named numeric vector of original parameter estimates (e.g. from run_nlme() results) to overlay as a solid vertical line on each panel. Names must match column names of bootstrap_results$parameter_estimates. Parameters not found are silently ignored. Pass NULL (default) to suppress.
ci: Numeric confidence level for the displayed credible interval (e.g. 0.95 draws the 2.5th and 97.5th percentile as dashed lines). Pass NULL to suppress CI lines.
parameters: Character vector of parameter names to include. NULL (default) includes all parameters.

Value

A ggplot2::ggplot() object. Use ggplot2::ggsave() to save it.

Examples

if (FALSE) { # \dontrun{
# After running bootstrap:
bs <- call_pharmpy_tool(
  id = "run1",
  model = model,
  tool = "bootstrap",
  options = list(samples = 200)
)

# Basic plot
plot_bootstrap(bs)

# With original estimates overlaid
orig <- c(POP_CL = 4.5, POP_V = 32.1, IIV_CL = 0.09)
p <- plot_bootstrap(bs, original_estimates = orig, ci = 0.95)
ggplot2::ggsave("bootstrap_params.png", p, width = 10, height = 8)
} # }