Summary
Fork lognormal populations, which differ only in one certain parameter Ϙ, the problem of finding the population with the largest value ofϑ is considered. For two-parameter lognormal families, several natural choices ofϑ are treated, where the problem can be solved, through logarithmic transformation of the observations, within the framework of estimating parameters ink, possibly restricted, normal populations. For three-parameter lognormal families, this standard approach of selecting in terms of natural estimators fails to work ifϑ is the “guaranteed lifetime”. For this case, a selection procedure is derived which is based on anL-statistic which has the smallest asymptotic variance. Of importance here is that it is location equivariant, whereas it does not matter what it actually estimates. Comparisons are made with other suitable selection rules, through the asymptotic relative efficiencies, as well as in an example of intermediate sample sizes. It is shown that only in the latter, the selection rule, which is based on the sample minima, compares favorably.
Similar content being viewed by others
References
Aitchison J, Brown JAC (1969) The lognormal distribution. 4th ed. Cambridge University Press, London
Cohen AC, Whitten BJ (1980) Estimation in the three-parameter lognormal distribution. J Amer Statist Assoc 75:399–404
Gibbons DI, McDonald GC (1975) Small-sample estimation for the lognormal distribution with unit shape parameter. IEEE Transactions on Reliability R-24:290–295
Gupta SS (1962) Life test sampling plans for normal and lognormal distributions. Technometrics 4:151–175
Gupta SS, McDonald GC, Galarneau DI (1974) Moments, product moments and percentage points of the order statistics from the lognormal distribution for samples of size twenty and less. Sankhyā Ser B 36:230–260
Gupta SS, Panchapakesan S (1979) Multiple decision procedures. J. Wiley, New York
Huber PJ (1981) Robust statistics. J. Wiley, New York
Hustý J (1981) Ranking and selection procedures for location parameter case based onL-estimates. Aplikace Matematiky 26:377–388
Hustý J (1984) Subset selection of the largest location parameter based onL-estimates. Aplikace Matematiky 29:397–410
Johnson NL, Kotz S (1970) Continuous univariate distributions-1. Houghton Mifflin, Boston
Kane VE (1982) Standard and goodness-of-fit parameter estimation methods for the three-parameter lognormal distribution. Commun Statist-Theor Meth 11:1935–1957
LaRiccia VN, Kindermann RP (1983) An asymptotically closed form estimator for the three-parameter lognormal distribution. Commun Statist-Theor Meth 12:243–261
Leadbetter MR, Lindgren G, Rootzén H (1983) Extremes and related properties of random sequences and processes. Springer-Verlag, New York
Lehmann EL (1983) Theory of point estimation. J Wiley, New York
Martz HF, Waller RA (1982) Bayesian reliability analysis. J. Wiley, New York
Müller PH, Neumann P, Storm R (1977) Tables for mathematical statistics (German). Carl-Hanser-Verlag, Munich
Miescke KJ (1979) Identification and selection procedures based on tests. Ann Statist 7:207–219
Sampford MR (1953) Some inequalities on Mill’s ratio and related functions. Ann Math Statist 24:130–132
Shorack GR (1972) Functions of order statistics. Ann Math Statist 43:412–427
Zellner A (1971) Bayesian and non-Bayesian analysis of the log-normal distribution and log-normal regression. J Amer Statist Assoc 66:327–330
Author information
Authors and Affiliations
Additional information
The research of this author was supported by the Office of Naval Research Contract N00014-88-K-0170 and NSF Grant Number DMS-8606964 at Purdue University. Reproduction in whole or in part is permitted for any purpose of the United States Government.
The research of this author was supported by the Air Force Office of Scientific Research Grant 85-0347 at the University of Illinois at Chicago.
Rights and permissions
About this article
Cite this article
Gupta, S.S., Miescke, K.J. On selecting the best ofk lognormal distributions. Metrika 36, 233–247 (1989). https://doi.org/10.1007/BF02614096
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02614096