This estimator estimates the real shape parameter based on generalized quantile plots based on mean excess functions, generalized median excesses or trimmed mean excesses.
Usage
shape.genquant(
xdat,
k,
type = c("genmean", "genmed", "trimmean"),
weight,
p = 0.5
)
Arguments
- xdat
n
vector of observations
- k
number of upper order statistics
- type
string indicating the estimator choice, one of genmean
, genmed
and trimmean
.
- weight
weight a kernel function on \([0,1]\)
- p
number between zero and one giving the proportion of order statistics for the second threshold
References
Beirlant, J., Vynckier P. and J.L. Teugels (1996). Excess functions and estimation of the extreme-value index. Bernoulli, 2(4), 293-318.