Extract Quantiles from 'density' Objects — quantile.density • mvMonitoring

Quantiles for objects of class density

# S3 method for density
quantile(x, probs = seq(0.25, 0.75, 0.25), names = TRUE, normalize = TRUE, ...)

Arguments

x: a object of class density or a list of densities
probs: numeric vector of probabilities with values in [0,1]. Note that elements very close to the boundaries return Inf or -Inf
names: logical; if TRUE, the result has a names attribute, resp. a rownames and colnames attributes. Set to FALSE for speedup with many probabilities
normalize: logical; if TRUE then the values in x$y are multiplied with a factor such that their integral is equal to one.
...: further arguments passed to or from other methods (currently unused)

Value

If x is of class density (or a list with exactly one element), a vector with quantiles. If x is a list of densities, then the output is a matrix of quantiles, with each matrix row corresponding to the respective density.

Details

This function is a near-exact copy of the quantile.density function from package BMS (https://CRAN.R-project.org/package=BMS). In spring of 2022, CRAN informed us that the BMS has been orphaned, so we copied the code (and corresponding documentation) we needed from it. See doi:10.18637/jss.v068.i04 for their paper.

The function quantile.density() applies generically to the built-in class density (as least for versions where there is no such method in the pre-configured packages). Note that this function relies on trapezoidal integration in order to compute the cumulative densities necessary for the calculation of quantiles.

Author

Stefan Zeugner, stefan.zeugner@ec.europa.eu

Martin Feldkircher, martin.feldkircher@da-vienna.ac.at

Examples

  rNorm_dens <- density(rnorm(100000))
  quantile(rNorm_dens)
#>          25%          50%          75% 
#> -0.676958638  0.002327997  0.680674963