Sankhya: The Indian Journal of Statistics

1995, Volume 57, Series A, Pt. 1, pp. 118--125

DETERMINANCY OF MEANS OF ORDER STATISTICS

By

 M. C. SPRUILL, Georgia Institute of Technology

SUMMARY.  If X1, X2,.... are independent random variables distributed as X and if $E[|X|] < \infty$ then the sequence $\rho_n = E[X_1 \vee X_2 \vee ... \vee X_n]$ of maximal moments determines the probability distribution of X. Exploiting the fact that a collection of functions is fundamental if and only if a corresponding moment problem is determinate yields simple proofs of the claim above, a theorem of Pollak on the determinacy of more general sequences of means of order statistics, and that a subsequence n(i) of the maximal moments determines the distribution if and only if $\sum_{i \geq 1} n^{-1}(i)=\infty$.

AMS (1990) subject classification.  Primary 62E10, secondary 41A10.

Key words and phrases. Fundamental sets, Müntz theorems, moments.

Full paper (PDF)

This article in mathematical reviews.