Estimation of Tails and Related Quantities Using the Number of Near-Extremes

From MaRDI portal





Motivated by an insurance context, the authors investigate the problem of estimating specific tail quantities of a common distribution function \(F\) in a random sample \(X_1,\dots,X_{N(t)},\) in which the claim sizes \(\{X_n\}\) are not necessarily assumed to be independent and \(\{N(t)\}\) is a counting process with non-decreasing sample paths. Based on the number of ``near-extremes of the \(X\)-sample, strong consistency as well as CLT type results are presented for both the tail index and the (finite) upper endpoint of a generalized Pareto distribution \(F,\) thus extending earlier work of the authors (2002) and Müller (2003).




Cited in
(25)








This page was built for publication: Estimation of Tails and Related Quantities Using the Number of Near-Extremes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4678815)