Lund University Publications

LUND UNIVERSITY LIBRARIES

Uniform Bounds on the Relative Error in the Approximation of Upper Quantiles for Sums of Arbitrary Independent Random Variables

and (2015) In Journal of Theoretical Probability 28(1). p.1-25
Abstract
Fix any n≥1. Let X~1,…,X~n be independent random variables. For each 1≤j≤n, X~j is transformed in a canonical manner into a random variable Xj. The Xj inherit independence from the X~j. Let sy and s∗y denote the upper 1y th −−− quantile of Sn=∑nj=1Xj and S∗n=sup1≤k≤nSk, respectively. We construct a computable quantity Q−−y based on the marginal distributions of X1,…,Xn to produce upper and lower bounds for sy and s∗y. We prove that for y≥8

6−1γ3y/16Q−−3y/16≤s∗y≤Q−−y

where

γy=12wy+1

and wy is the unique solution of

(wyeln(yy−2))wy=2y−4

for wy>ln(yy−2), and for y≥37

19γu(y)Q−−u(y)<sy≤Q−−y

where

u(y)=3y32(1+1−643y−−−−−−√).

The distribution of Sn is... (More)
Fix any n≥1. Let X~1,…,X~n be independent random variables. For each 1≤j≤n, X~j is transformed in a canonical manner into a random variable Xj. The Xj inherit independence from the X~j. Let sy and s∗y denote the upper 1y th −−− quantile of Sn=∑nj=1Xj and S∗n=sup1≤k≤nSk, respectively. We construct a computable quantity Q−−y based on the marginal distributions of X1,…,Xn to produce upper and lower bounds for sy and s∗y. We prove that for y≥8

6−1γ3y/16Q−−3y/16≤s∗y≤Q−−y

where

γy=12wy+1

and wy is the unique solution of

(wyeln(yy−2))wy=2y−4

for wy>ln(yy−2), and for y≥37

19γu(y)Q−−u(y)<sy≤Q−−y

where

u(y)=3y32(1+1−643y−−−−−−√).

The distribution of Sn is approximately centered around zero in that P(Sn≥0)≥118 and P(Sn≤0)≥165. The results extend to n=∞ if and only if for some (hence all) a>0

∑j=1∞E{(X~j−mj)2∧a2}<∞. (Less)
author
and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
quantile approximation, tail probabilities, Sum of independent random variables, tail distributions, Hofmann-J/orgensen/Klass- Nowicki Inequality
in
Journal of Theoretical Probability
volume
28
issue
1
pages
1 - 25
publisher
Springer
external identifiers
• scopus:84929848604
• wos:000387922500008
ISSN
1572-9230
DOI
10.1007/s10959-015-0615-y
language
English
LU publication?
yes
id
dea1ba64-9012-4897-ba69-32542c776a45 (old id 1971437)
2016-04-01 13:29:02
date last changed
2022-01-27 19:25:27
```@article{dea1ba64-9012-4897-ba69-32542c776a45,
abstract     = {{Fix any n≥1. Let X~1,…,X~n be independent random variables. For each 1≤j≤n, X~j is transformed in a canonical manner into a random variable Xj. The Xj inherit independence from the X~j. Let sy and s∗y denote the upper 1y th −−− quantile of Sn=∑nj=1Xj and S∗n=sup1≤k≤nSk, respectively. We construct a computable quantity Q−−y based on the marginal distributions of X1,…,Xn to produce upper and lower bounds for sy and s∗y. We prove that for y≥8<br/><br>
6−1γ3y/16Q−−3y/16≤s∗y≤Q−−y<br/><br>
where<br/><br>
γy=12wy+1<br/><br>
and wy is the unique solution of<br/><br>
(wyeln(yy−2))wy=2y−4<br/><br>
for wy&gt;ln(yy−2), and for y≥37<br/><br>
19γu(y)Q−−u(y)&lt;sy≤Q−−y<br/><br>
where<br/><br>
u(y)=3y32(1+1−643y−−−−−−√).<br/><br>
The distribution of Sn is approximately centered around zero in that P(Sn≥0)≥118 and P(Sn≤0)≥165. The results extend to n=∞ if and only if for some (hence all) a&gt;0<br/><br>
∑j=1∞E{(X~j−mj)2∧a2}&lt;∞.}},
author       = {{Klass, Michael J and Nowicki, Krzysztof}},
issn         = {{1572-9230}},
keywords     = {{quantile approximation; tail probabilities; Sum of independent random variables; tail distributions; Hofmann-J/orgensen/Klass- Nowicki Inequality}},
language     = {{eng}},
number       = {{1}},
pages        = {{1--25}},
publisher    = {{Springer}},
series       = {{Journal of Theoretical Probability}},
title        = {{Uniform Bounds on the Relative Error in the Approximation of Upper Quantiles for Sums of Arbitrary Independent Random Variables}},
url          = {{http://dx.doi.org/10.1007/s10959-015-0615-y}},
doi          = {{10.1007/s10959-015-0615-y}},
volume       = {{28}},
year         = {{2015}},
}

```