Abstract
Let us consider k(≥ 2) independent random variables U1, . . . ,Uk where Ui is distributed as the Student's t random variable with a degree of freedom mi, i=1, . . . ,k. Here, m1, . . . ,mk are arbitrary positive integers. We denote m=(m1, . . . ,mk) and Uk:k=max {U1, . . . ,Uk}, the largest Student's t random variable. Having fixed 0<α <1, let a≡ a(k,α) and hm ≡ hm (k,α) be two positive numbers for which we can claim that (i) Φk(a)−Φk(−a)=1−α, and (ii) P{−hm≤ Uk:k≤ hm}=1−α. Then, we proceed to derive a Cornish–Fisher expansion (Theorem 3.1) of the percentage point hm. This expansion involves “a” as well as expressions such as Σi=1 k mi −1, Σi=1 kmi −2, and Σi=1 k mi −3. The corresponding approximation of hm is shown to be remarkably accurate even when k or m1, . . . ,mk are not very large.
Similar content being viewed by others
References
M. Aoshima and N. Mukhopadhyay, “Two-stage estimation of a linear function of normal means with second-order approximations,” Sequential Anal. vol. 21 pp. 109–144, 2002.
H. J. Chen and E. J. Dudewicz, “Procedures for fixed-width interval estimation of the largest normal mean,” J. Amer. Statist. Assoc. vol. 71 pp. 752–756, 1976.
S. S. Gupta and S. Panchapakesan, Multiple Decision Procedures, Wiley, Inc.: New York, 1979.
N. L. Johnson and S. Kotz, Continuous Univariate Distributions, 2. Wiley, Inc.: New York, 1970.
L. Kuo and N. Mukhopadhyay, “Multi-stage point and interval estimation of the largest mean of k normal populations and the associated second-order properties,” Metrika vol. 37 pp. 291–300, 1990.
N. Mukhopadhyay and M. Aoshima, “Percentage points of the record Student's t random variable,” Statist. Tech. Rep. #02–30, Department of Statistics, University of Connecticut, Storrs, 2002.
N. Mukhopadhyay, S. Chattopadhyay, and S. K. Sahu, “Further developments in estimation of the largest mean of k normal populations,” Metrika vol. 40 pp. 173–183, 1993.
N. Mukhopadhyay and T. K. S. Solanky, Multistage Selection and Ranking Procedures, Marcel Dekker, Inc.: New York, 1994.
K. M. L. Saxena, “A single-sample procedure for the estimation of the largest mean,” J. Amer. Statist. Assoc. vol. 71 pp. 147–148, 1976.
K. M. L. Saxena and Y. L. Tong, “Interval estimation of the largest mean of k normal populations with known variance,” J. Amer. Statist. Assoc. vol. 64 pp. 296–299, 1969.
Y. L. Tong, “Multi-stage interval estimations of the largest mean of k normal populations,” J. Roy. Statist. Soc., Ser. B vol. 32 pp. 272–277, 1970.
Y. L. Tong, “An asymptotically optimal sequential procedure for the estimation of the largest mean,” Ann. Statist. vol. 1 pp. 175–179, 1973.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Mukhopadhyay, N., Aoshima, M. Percentage Points of the Largest Among Student's T Random Variable. Methodology and Computing in Applied Probability 6, 161–179 (2004). https://doi.org/10.1023/B:MCAP.0000017711.83727.a5
Issue Date:
DOI: https://doi.org/10.1023/B:MCAP.0000017711.83727.a5