EDGEWORTH EXPANSIONS FOR -STATISTICS
Ivo B. Alberink
Gyula Pap
Martien C. A. van Zuijlen
Abstract: We study the approximation by a short Edgeworth expansion of the distribution
function of normalized linear combinations
of
order statistics of
independent random variables with common distribution function
Under the assumptions
|cjn| | < Cn-p1
1 --p2
, | |
|
|cjn - cj-1,n| | < Cn-q1
1 --q2
, | |
|
|cj+1,n - 2cjn + cj-1,n| | < Cn-r1
1 --r2
, | |
|
(F-1)'(s) | < C[s(1 - s)]- | | |
for some
with an appropriate
balance in these parameters, and under additional moment conditions, the rate of
uniform convergence is shown to be of order
Moreover, a special case is
considered where the
are generated by a sequence of weight functions of a special
structure.
1991 AMS Mathematics Subject Classification: 62E20.
Key words and phrases: Linear combinations of order statistics, Edgeworth expansions,
rate of convergence.