qreg: Estimate Quantile Region
In perej1/extreme: Multivariate Extreme Region Estimation for Elliptical Distributions

qreg	R Documentation

Estimate Quantile Region

Description

Estimates (1-p)-quantile region.

Usage

qreg(
  x,
  p,
  method = "sample",
  qmethod = "sample",
  k = NULL,
  alpha = NULL,
  mu = NULL,
  sigma = NULL
)

Arguments

`x`	A matrix representing sample from elliptical distribution. Rows are observations and columns dimensions.
`p`	Probability corresponding to `(1-p)`-quantile. Values in interval `(0, 1)` are accepted.
`method`	How location and scatter should be estimated?
`qmethod`	How quantile should be estimated?
`k`	Threshold parameter for the estimator. Threshold parameter has to be in set `{1,2,\ldots,n}` where `n` is equal to `nrow(x)`.
`alpha`	Controls size of the subsets over which the determinant is minimized when `method = "mcd"`.
`mu`	Given estimate for location.
`sigma`	Given estimate for scatter.

Details

Two estimators are provided. Extreme quantile estimator is consistent only for heavy-tailed elliptical distributions. Extreme value index is estimated with separating Hill estimator. Other estimator is the conventional sample quantile.