English

Second Order Approximations for Slightly Trimmed Sums

Probability 2014-10-21 v1

Abstract

We investigate the second order asymptotic behavior of trimmed sums Tn=1ni=\kn+1n\mn\xinT_n=\frac 1n \sum_{i=\kn+1}^{n-\mn}\xin, where \kn\kn, \mn\mn are sequences of integers, 0\kn<n\mnn0\le \kn < n-\mn \le n, such that min(\kn,\mn)\min(\kn, \mn) \to \infty, as \nty\nty, the \xin\xin's denote the order statistics corresponding to a sample X1,...,XnX_1,...,X_n of nn i.i.d. random variables. In particular, we focus on the case of slightly trimmed sums with vanishing trimming percentages, i.e. we assume that max(\kn,\mn)/n0\max(\kn,\mn)/n\to 0, as \nty\nty, and heavy tailed distribution FF, i.e. the common distribution of the observations FF is supposed to have an infinite variance. We derive optimal bounds of Berry -- Esseen type of the order O(rn1/2)O\bigl(r_n^{-1/2}\bigr), rn=min(\kn,\mn)r_n=\min(\kn,\mn), for the normal approximation to TnT_n and, in addition, establish one-term expansions of the Edgeworth type for slightly trimmed sums and their studentized versions. Our results supplement previous work on first order approximations for slightly trimmed sums by Csorgo, Haeusler and Mason (1988) and on second order approximations for (Studentized) trimmed sums with fixed trimming percentages by Gribkova and Helmers (2006, 2007).

Keywords

Cite

@article{arxiv.1104.3347,
  title  = {Second Order Approximations for Slightly Trimmed Sums},
  author = {N. V. Gribkova and R. Helmers},
  journal= {arXiv preprint arXiv:1104.3347},
  year   = {2014}
}

Comments

37 pages, to appear in Theory Probab. Appl

R2 v1 2026-06-21T17:55:17.683Z