## Abstract Let {__S~n~__, __n__ β₯ 1} be partial sums of independent identically distributed random variables. The almost sure version of CLT is generalized on the case of randomly indexed sums {__S~Nn~__, __n__ β₯ 1}, where {__N~n~__, __n__ β₯ 1} is a sequence of positive integerβvalued random varia
On weak and Moments Convergence of Randomly Indexed Sums
β Scribed by A. Krajka; Z. Rychlik
- Publisher
- John Wiley and Sons
- Year
- 1992
- Tongue
- English
- Weight
- 465 KB
- Volume
- 157
- Category
- Article
- ISSN
- 0025-584X
No coin nor oath required. For personal study only.
β¦ Synopsis
Abstract
Let {X~n~, n β©Ύ 1) be a sequence of independent random variables such that EX~n~ = a~n~, E(X~n~ β a~n~)^2^ = Ο, n β©Ύ 1. Let {N~n~, n β©Ύ 1} be a sequence of positive integerβvalued random variables. Let us put
In this paper we present necessary and sufficient conditions for weak and moments convergence of the sequence {(S__βL~n~__)/s~n~, n β©Ύ 1}, as n β β. Hermite polinomial type limit theorems are also considered. The obtained results extend the main theorem of M. Finkelstein and H. G. Tucker (1989).
π SIMILAR VOLUMES
For each n, let ( S n k ) , 1 S k s k,, be a mean zero square -integrable martingale adapted to increasing a-fields ($nk), O s k s h n , and let ( b n k ) , OSkaE,, be a system of random variables such that bno=O -=bnl-=... -= bnkn= 1 and such that bnk is Snn,k-l measurable for each k. We present su