GET (version 1.0-5)

plot.combined_global_envelope2d: Plotting function for combined 2d global envelopes

Description

If fixedscales is FALSE (or 0) all images will have separate scale. If fixedscales is TRUE (or 1) each x[[i]] will have a common scale. If fixedscales is 2 all images will have common scale.

If more than one envelope has been calculated (corresponding to several coverage/alpha), only the largest one is plotted.

Usage

# S3 method for combined_global_envelope2d
plot(
  x,
  fixedscales = 2,
  labels,
  what = c("obs.sign", "obs", "lo", "hi", "lo.sign", "hi.sign"),
  sign.type = c("circles", "contour", "col"),
  sign.col = c("blue", "red"),
  transparency = 155/255,
  digits = 3,
  ...
)

Arguments

x

A 'global_envelope' object for two-dimensional functions

fixedscales

0, 1 or 2. See details.

labels

A character vector of suitable length giving the labels for the separate plots. Default exists. This parameter allows replacing the default.

what

Character vector specifying what information should be plotted for 2d functions. A combination of: Observed ("obs"), upper envelope ("hi"), lower envelope ("lo"), observed with significantly higher values highlighted ("hi.sign"), observed with significantly lower values highlighted ("lo.sign"), observed with significantly (lower and higher) values highlighted ("obs.sign"). Default to the last one. Combination c("obs", "lo", "hi", "lo.sign", "hi.sign") can also be of interest (earlier default).

sign.type

Either "col" for color showing the significant region, or "contour" for colored contour showing the significant region, or "circles" for plotting circles at locations where the observed function exceeds the envelope: diameters proportional to (obs-hi)/(hi-lo) for values above the envelope and (lo-obs)/(hi-lo) for values below the envelope. In the one-sided (testing) case, the divisors are instead (hi-central) (case 'greater') and (central-lo) (case 'less'). Default to "circles".

sign.col

A vector of length two giving the colors for significant parts below the envelope (first value) and above the envelope (second value).

transparency

A number between 0 and 1 (default 155/255, 60 Similar to alpha of rgb. Used in plotting the significant regions for 2d functions.

digits

The number of digits used for printing the p-value or p-interval in the default main.

...

Ignored.

Examples

data("abide_9002_23")
iset <- subset(abide_9002_23[['curve_set']], 1:50)
factors <- abide_9002_23[['factors']][1:50,]
# \dontshow{
# Cut the data to reduce time
iset$r <- iset$r[1:7,]
iset$funcs <- iset$funcs[1:7, ]
# }
res <- graph.flm(nsim = 19, # Increase nsim for serious analysis!
  formula.full = Y ~ Group + Sex + Age,
  formula.reduced = Y ~ Sex + Age,
  curve_sets = list(Y=iset), factors = factors,
  contrasts = FALSE, GET.args = list(type="area"))
plot(res)
# \donttest{
plot(res, what=c("obs", "hi"))

plot(res, what=c("hi", "lo"), fixedscales=1)

plot(res, what=c("obs", "lo", "hi"), fixedscales=FALSE)

if(requireNamespace("gridExtra", quietly=TRUE)) {
  # Edit style of "fixedscales = 2" plots
  plot(res, what=c("obs", "hi")) + ggplot2::theme_minimal()
  plot(res, what=c("obs", "hi")) + ggplot2::theme_bw()

  # Edit style (e.g. theme) of "fixedscales = 1 or 0" plots
  gs <- lapply(res, function(x, what) { plot(x, what=what) +
     ggplot2::ggtitle("") }, what=c("obs", "hi"))
  gridExtra::grid.arrange(grobs=gs, ncol=1, top="My main")

  gs <- outer(res, c("obs", "hi"), FUN=Vectorize(function(res, what)
    list(plot(res, what=what) + ggplot2::ggtitle("") +
      ggplot2::theme(axis.ticks=ggplot2::element_blank(),
      axis.text=ggplot2::element_blank(), axis.title=ggplot2::element_blank()))))
  gridExtra::grid.arrange(grobs=t(gs))
}
# }