[Deprecated function]
Given a random sample of n exceedances, the estimator
returns an estimator of the shape parameter or extreme
value index using a kernel of order 3, based on
k largest exceedances of xdat. Oorschot et al. (2023) parametrize the model in terms of \(m=n-k\), so that \(m=n\) corresponds to using only the three largest observations.
PickandsXU(xdat, m)vector of observations of length \(n\)
number of largest order statistics \(3 \leq k \leq n\).
The calculations are based on the recursions provided in Lemma 4.3 of Oorschot et al.
Oorschot, J, J. Segers and C. Zhou (2023), Tail inference using extreme U-statistics, Electronic Journal of Statistics, 17(1): 1113-1159. tools:::Rd_expr_doi("10.1214/23-EJS2129")