Plot pairs of quantiles

Utility function to plot pairs of quantiles obtained with Pr().

Usage

plotquantiles(
  x,
  y,
  xdomain = NULL,
  alpha.f = 0.25,
  col = palette(),
  border = NA,
  type = "n",
  ...
)

Arguments

x

Numeric or character: vector of x-coordinates. See flexiplot().

y

Numeric: a matrix having as many rows as x and an even number of columns, with one column per quantile. Typically these quantiles have been obtained with Pr(), as their $quantiles value. This value is a three-dimensional array, and one of its columns (corresponding to the possible values of the X argument of Pr()) or one of its rows (corresponding to the possible values of the Y argument of Pr()) should be selected before being used as y input.

xdomain

Character or numeric or NULL (default): vector of possible values of the variable represented in the x-axis, if the x argument is a character vector. The ordering of the values is respected. If NULL, then unique(x) is used.

alpha.f

Numeric, default 0.25: opacity of the quantile bands, 0 being completely invisible and 1 completely opaque.

col

Fill colour of the quantile bands. Can be specified in any of the usual ways, see for instance grDevices::col2rgb(). Default #4477AA.

border

Fill colour of the quantile bands. Can be specified in any of the usual ways, see for instance grDevices::col2rgb(). If NA (default), no border is drawn.

type

see analogous argument in flexiplot().

...

Other parameters to be passed to flexiplot().

Value

NULL, invisibly; produces a plot, see graphics::matplot().

See also

Pr() to calculate posterior probabilities and quantiles.

plot.probability() to directly plot posterior probabilities and quantiles contained in a probability object.

flexiplot() for more general plots.

Examples

## Load the example `learnt` object calculated from the "penguins" dataset;
## variates: 'species' and 'bill_len'
learnt <- learntExample

## create a grid of values for variate "bill length",
## based on the information in the dataset and metadata:
values <- vrtgrid(vrt = 'bill_len', learnt = learnt)

## calculate the probabilities and quantiles
probs <- Pr(Y = data.frame(bill_len = values), learnt = learnt, parallel = 1)

## plot the quantiles, setting lower plot range to zero
plotquantiles(x = values, y = probs$quantiles[, 1, ], ylim = c(0, NA),
  xlab = 'bill length', ylab = 'probability')

## add a plot of the probabilities in thick dashed red
flexiplot(x = values, y = probs$values, lwd = 5, lty = 2, col = 2, add = TRUE)