Plot a legend for a visual predictive check (VPC)

Usage

plot_vpc_legend(
  ci = 0.9,
  pi = c(0.05, 0.95),
  shown = NULL,
  loq = NULL,
  show_loq_value = TRUE,
  style = NULL,
  type = c("cont", "cens"),
  ...,
  lloq = NULL,
  stats = NULL
)

Arguments

ci: Numeric confidence level for simulation intervals (e.g., 0.90 for 90% CI). Should match argument passed to plot_vpc_cont(). Default is 0.90.
pi: prediction intervals plotted. Should match argument passed to plot_vpc_cont(). Default is c(0.05, 0.95). Ignored when type = "cens" (cens VPCs do not have prediction intervals).
shown: Layer visibility settings created by plot_vpc_shown(). Defaults can be viewed by running plot_vpc_shown() with no arguments.
loq: Numeric scalar or vector of LLOQ values to label in the legend, or NULL to omit. One legend entry is drawn at the highest value, matching the single LLOQ line per panel, with the style's loq_line linetype and linewidth. Pass stats$config$loq from a df_vpcstats() result to mirror the reference lines drawn by plot_build_vpc().
show_loq_value: Logical. If TRUE (default), the LLOQ entry is labeled "LLOQ = <value>" with the highest loq value; if FALSE, it is labeled "LLOQ". Ignored when loq is NULL.
style: A ggstylekit::style_spec() controlling legend aesthetics. Defaults to style_vpc(); view the defaults by running style_vpc() with no arguments. Should match the style passed to plot_vpc_cont(). The per-series maps supply each key's color, fill, shape, linetype, and linewidth; legend.title.position and legend.title.hjust place the legend titles. Other style fields do not apply to the legend panel.
type: One of "cont" (default) or "cens". Selects the labels and layer set the legend describes. Under "cont", the central-tendency entries are labeled "Obs Med", "Sim Med", and "Sim <ci>% CI Med" and the pi entries ("Obs <p1>th and <p2>th", "Sim <p1>th - <p2>th", "Sim <ci>% CI <p1>th and <p2>th") are included when their shown keys are on. Under "cens", the three central-tendency labels are relabeled to "Obs Prop BLQ", "Sim Prop BLQ", "Sim <ci>% CI Prop BLQ" and all pi-related entries are suppressed regardless of shown (cens VPCs have no prediction interval).
...: Other arguments passed to ggplot2::theme(), applied after the style so they override it.
lloq: Use loq. Do not pass both.
stats: A df_vpcstats() result, or NULL (default). When given, ci, pi, and loq default to stats$config$ci, stats$config$pi, and stats$config$loq, so the legend matches a plot built from the same container; any of them passed explicitly overrides the container (loq = NULL omits the LLOQ entry). type, shown, and style are not stored in the container; pass the ones used for the plot.

Value

a ggplot2 object

Examples

plot_vpc_legend()

plot_vpc_legend(
pi = c(0.025, 0.975),
ci = 0.95,
 shown = plot_vpc_shown(obs_point = FALSE, obs_pi_line = TRUE,
 sim_pi_line = FALSE, sim_pi_area = FALSE, sim_pi_ci = TRUE,
 obs_median_line = TRUE,
 sim_median_line = FALSE, sim_median_ci = TRUE))

## Cens VPC legend
plot_vpc_legend(type = "cens", loq = 1)

## LLOQ entry without its value
plot_vpc_legend(loq = c(0.5, 1), show_loq_value = FALSE)

## ci, pi, and loq from a df_vpcstats() result
simout <- df_mrgsim_replicate(data = dplyr::filter(data_sad, CMT %in% c(1, 2)),
                             model = model_mread_load(model = "pkmodel"),
                             replicates = 10, dv_var = ODV)
#> Loading model from cache.
out <- df_vpcstats(simout, loq = 1, pi = c(0.1, 0.9))
plot_vpc_legend(stats = out)

Plot a legend for a visual predictive check (VPC)

Usage

Arguments

Value

See also

Examples