Plots VPC ribbons (a confidence interval around the median, lower and upper
percentiles) from a std_campsis_tbl, i.e. the default output of
simulate() with multiple replicates and no custom outfun
(DefaultOutfun). In this case the data contains all
unsummarised individual profiles, so the per-replicate prediction interval
is computed first (the lower/median/upper percentiles of variable
across individuals within each replicate), before the confidence interval
around each percentile is computed across replicates.
Usage
# S3 method for class 'std_campsis_tbl'
vpc_plot(
x,
variable = "CONC",
strata = "auto",
pi_level = 0.9,
ci_level = 0.9,
alpha = 0.15,
facet = TRUE,
...
)Arguments
- x
a
std_campsis_tblobject with areplicatecolumn- variable
name of the column to summarise on the y-axis. Defaults to
"CONC". An informative error is raised when the column is absent.- strata
stratification for the VPC. One of
"auto",NULL, or a named character vector of length 1. See Details.- pi_level
prediction interval level, i.e. the percentiles of the data computed within each replicate. Default is
0.90(90% PI, i.e. the 5th, 50th and 95th percentiles).- ci_level
confidence interval level for the ribbons around each percentile across replicates. Default is
0.90(90% CI).- alpha
transparency of the ribbons. Default is
0.15.- facet
how the stratification variable is displayed when present.
TRUE(default) draws one panel per stratum;FALSEoverlays the strata in a single panel and maps the stratum to the ribbon fill colour. Forwarded tovpcPlot.- ...
additional arguments (currently unused, reserved for future use)
Details
Stratification is resolved as follows:
"auto"(default):ARMis used when more than one distinct arm is detected; otherwiseSCENARIOis used when more than one distinct scenario is detected; otherwiseNULL(no stratification). Only one stratification variable is applied becausevpcPlotsupports at most one.NULL: no stratification.named character vector: passed directly to
vpcPlot(e.g.c(ARM = "all")orc(SCENARIO = "all")).