On numbers of observations near randomly indexed order statistics (Q2267607): Difference between revisions

Generalizing work of \textit{A.G. Pakes} and \textit{F.W. Steutel} [Aust. J. Stat. 39, No. 2, 179--192 (1997; Zbl 0908.60040)], and others, the author considers the numbers of observations of a random sample that fall in the neighborhoods of several order statistics. She obtains limit theorems for the multivariate distributions of these quantities, both for samples indexed by a nonrandom variable \(n\), \(n \rightarrow \infty\), and for samples indexed by a process \(N(t)\), \(t \rightarrow \infty\). It is shown that these numbers are often asymptotically independent, and have Poisson type distributions. The mathematics is straightforward, but rather complicated.