An upper bound for the total variation distance between the distribution of the sum of a sequence of r.v.'s and that of a compound Poisson is derived. Its applications to a general independent sequence and Markov-binomial sequence are demonstrated. (~
Remarks on compound Poisson approximation of Gaussian random sequences
✍ Scribed by Enkelejd Hashorva; Jürg Hüsler
- Publisher
- Elsevier Science
- Year
- 2002
- Tongue
- English
- Weight
- 117 KB
- Volume
- 57
- Category
- Article
- ISSN
- 0167-7152
No coin nor oath required. For personal study only.
✦ Synopsis
Let {X i ; i ¿ 1} be a sequence of m-dependent stationary standard Gaussian random variables and u n ; n ∈ N some positive constants. In this note we generalise results of Raab (Extremes 1(3) (1999) 29.), who considered compound Poisson approximation for W n = n i=1 1{X i ¿ u n } the number of exceedances above the level u n . More precisely, the main result concerns an upper asymptotic bound for the total variational distance d TV (W n ; CP( * )) where CP( * ) d = N 1 + 2N 2 + • • • + rN r ; with 2 6 r 6 2m and N i d = Poi( i ); i ¿ 0 are independent Poisson random variables.
📜 SIMILAR VOLUMES
We present multivariate generalizations of some classical results on the accuracy of Poisson approximation for the distribution of a sum of 0 -1 random variables. A multivariate generalization of Bradley's theorem (Michigan Math. J. 30 (1983) 69) is established as well.