VineCopula (version 2.6.0)

pairs.copuladata: Pairs Plot of Copula Data

Description

This function provides pair plots for copula data. Using default setting it plots bivariate contour plots on the lower panel, scatter plots and correlations on the upper panel and histograms on the diagonal panel.

Usage

# S3 method for copuladata
pairs(
  x,
  labels = names(x),
  ...,
  lower.panel = lp.copuladata,
  upper.panel = up.copuladata,
  diag.panel = dp.copuladata,
  label.pos = 0.85,
  cex.labels = 1,
  gap = 0,
  method = "kendall",
  ccols = terrain.colors(11),
  margins = "norm"
)

Arguments

x

copuladata object.

labels

variable names/labels.

...

other graphical parameters (see graphics::par()) or options passed to BiCopKDE().

lower.panel

panel function to be used on the lower diagonal panels (if not supplied, a default function is used)

upper.panel

panel function to be used on the upper diagonal panels (if not supplied, a default function is used)

diag.panel

panel function to be used on the diagonal panels (if not supplied, a default function is used)

label.pos

y position of labels in the diagonal panel; default: label.pos = 0.85.

cex.labels

magnification to be used for the labels of the diagonal panel; default: cex.labels = 1.

gap

distance between subplots, in margin lines; default: gap = 0.

method

a character string indicating which correlation coefficients are computed. One of "pearson", "kendall" (default), or "spearman"

ccols

color to be used for the contour plots; default: ccols = terrain.colors(30).

margins

character; margins for the contour plots. Options are:
"unif" for the original copula density, "norm" for the transformed density with standard normal margins, "exp" with standard exponential margins, and "flexp" with flipped exponential margins.

Author

Tobias Erhardt

See Also

graphics::pairs(), as.copuladata(), BiCopKDE()

Examples

data(daxreturns)
daxreturns <- daxreturns[1:50, ]
data <- as.copuladata(daxreturns)
sel <- c(4,5,14,15)

## pairs plot with default settings
pairs(data[sel])

## pairs plot with custom settings
nlevels <- 20
pairs(data[sel], cex = 2, pch = 1, col = "black",
      diag.panel = NULL, label.pos = 0.5,
      cex.labels = 2.5, gap = 1,
      method = "pearson", ccols = heat.colors(nlevels),
      margins = "flexp")

## pairs plot with own panel functions
up <- function(x, y) {
  # upper panel: empirical contour plot
  op <- par(usr = c(-3, 3, -3, 3), new = TRUE)
  BiCopKDE(x, y,
           levels = c(0.01, 0.05, 0.1, 0.15, 0.2),
           margins = "exp",
           axes = FALSE)
  on.exit(par(op))
}

lp <- function(x, y) {
  # lower panel: scatter plot (copula data) and correlation
  op <- par(usr = c(0, 1, 0, 1), new = TRUE)
  points(x, y, pch = 1, col = "black")
  r <- cor(x, y, method = "spearman") # Spearman's rho
  txt <- format(x = r, digits = 3, nsmall = 3)[1]
  text(x = 0.5, y = 0.5, labels = txt, cex = 1 + abs(r) * 2, col = "blue")
  on.exit(par(op))
}

dp <- function(x) {
  # diagonal panel: histograms (copula data)
  op <- par(usr = c(0, 1, 0, 1.5), new = TRUE)
  hist(x, freq = FALSE, add = TRUE, col = "brown", border = "black", main = "")
  abline(h = 1, col = "black", lty = 2)
  on.exit(par(op))
}

nlevels <- 20
pairs(data[sel],
      lower.panel = lp, upper.panel = up, diag.panel = dp, gap = 0.5)