Electronic Journal of Statistics

On the Nile problem by Sir Ronald Fisher

Abram M. Kagan and Yaakov Malinovsky

Full-text: Open access

Abstract

The Nile problem by Ronald Fisher may be interpreted as the problem of making statistical inference for a special curved exponential family when the minimal sufficient statistic is incomplete. The problem itself and its versions for general curved exponential families pose a mathematical-statistical challenge: studying the subalgebras of ancillary statistics within the $\sigma$-algebra of the (incomplete) minimal sufficient statistics and closely related questions of the structure of UMVUEs.

In this paper a new method is developed that, in particular, proves that in the classical Nile problem no statistic subject to mild natural conditions is a UMVUE. The method almost solves an old problem of the existence of UMVUEs. The method is purely statistical (vs. analytical) and works for any family possessing an ancillary statistic. It complements an analytical method that uses only the first order ancillarity (and thus works when the existence of ancillary subalgebras is an open problem) and works for curved exponential families with polynomial constraints on the canonical parameters of which the Nile problem is a special case.

Article information

Source
Electron. J. Statist., Volume 7 (2013), 1968-1982.

Dates
First available in Project Euclid: 18 July 2013

Permanent link to this document
https://projecteuclid.org/euclid.ejs/1374153370

Digital Object Identifier
doi:10.1214/13-EJS832

Mathematical Reviews number (MathSciNet)
MR3084678

Zentralblatt MATH identifier
1293.62009

Subjects
Primary: 62B05: Sufficient statistics and fields
Secondary: 62F10: Point estimation

Keywords
Ancillarity complete sufficient statistics curved exponential families UMVUEs

Citation

Kagan, Abram M.; Malinovsky, Yaakov. On the Nile problem by Sir Ronald Fisher. Electron. J. Statist. 7 (2013), 1968--1982. doi:10.1214/13-EJS832. https://projecteuclid.org/euclid.ejs/1374153370


Export citation

References

  • [1] Akhiezer, N. I. (1965)., The classical moment problem and some related questions in analysis. Hafner Publishing Co., New York.
  • [2] Bahadur, R. R. (1957). On unbiased estimates of uniformly minimum variance., Sankhya Ser. A 18, 211–224.
  • [3] Barra, J-R. (1971)., Notions Fondamentales de Statistique Mathématique. Dunod, Paris. (English Translation (1981): Mathematical Basis of Statistics. Academic Press, New York).
  • [4] Basu, D. (1955). On statistics independent of a complete sufficient statistic., Sankhya Ser. A 15, 377–380.
  • [5] Basu, D. (1958). On statistics independent of sufficient statistics., Sankhya Ser. A 18, 223–226.
  • [6] Bondesson, L. (1975). Uniformly minimum variance estimation in location parameter families., Ann. Statist. 3, 637–600.
  • [7] Bondesson, L. (1983). On uniformly minimum variance unbiased estimation when no complete sufficient statistics exist., Metrika 30, 49–54.
  • [8] Casella, G. and Berger, R. L. (2002)., Statistical Inference, 2nd ed. Duxbury.
  • [9] Ferguson, T. S. (1964). A characterization of the exponential distribution., Ann. Math. Statist. 35, 1199–1207.
  • [10] Ferguson, T. S. (1967). On characterizing distributions by properties of order statistics., Sankhya Ser. A 29, 265–277.
  • [11] Fisher, R. A. (1936). Uncertain inference., Proceedings of the American Academy of Arts and Sciences 71, 245–258.
  • [12] Fisher, R. A. (1973)., Statistical Methods and Scientific Inference, 3rd ed. Hafner Press.
  • [13] Flatto, F. and Shepp, L. (1990). Problem of the Nile., SIAM Review 32, 302–304.
  • [14] Gather, U. (1996). Characterizing distributions by properties of order statistics – a partial review., H.N. Nagaraja, P.K. Sen, D.F. Morrison (Eds.), Statistical Theory and Applications, 89–103. Springer, New York.
  • [15] Gleser, L. J. and Healy, J. D. (1976). Estimating the mean of a normal distribution with known coefficient of variation., J. Amer. Statist. Assoc. 71, 977–981.
  • [16] Hodges, J. L., Jr. and Lehmann, E. L. (1951). Some applications of the Cramér-Rao inequality., Proceedings of the Second Berkeley Symposium on Mathematical Statistics and Probability, 1950, 13–22. University of California Press, Berkeley and Los Angeles.
  • [17] Hinkley, D. V. (1977). Conditional inference about a normal mean with known coefficient of variation., Biometrika 64, 105–108.
  • [18] Kagan, A. M., Linnik, Yu. V. and Rao, C. R. (1965). On a characterization of the normal law based on a property of the sample average., Sankhya Ser. A 27, 405–406.
  • [19] Kagan, A. M. (1966). Two remarks on characterization of sufficiency (Russian)., Limit Theorems Statist. Inference, Izdat. “Fan”, Tashkent, 60–66.
  • [20] Kagan, A. M. and Palamodov, V. P. (1967a). Incomplete exponential families and minimum variance unbiased estimators, I., Theor. Probab. Appl. 12, 36–46.
  • [21] Kagan, A. M. and Palamodov, V. P. (1967b). Conditions of optimal unbiased estimation of parametric functions for incomplete exponential families with polynomial constraints., Dokl. Akad. Nauk SSSR 175, 1216–1218.
  • [22] Kagan, A. M. and Palamodov, V. P. (1968). New results in the theory of estimation and testing hypotheses for problems with nuisance parameters., Supplement to Linnik’s “Statistical Problems with Nuisance Parameters”. American Mathematical Society, Providence.
  • [23] Kagan, A. M., Linnik, Yu. V. and Rao, C. R. (1973)., Characterization Problems in Mathematical Statistics. Wiley, New York.
  • [24] Kagan, A. M. (2002). Sufficiency and ancillarity in characterization problems., J. Statist. Plann. Inference 102, 223–228.
  • [25] Kagan, A. M. and Rao, C. R. (2006). On estimation of a location parameter in presence of an ancillary component., Theor. Probab. Appl. 50, 172–176.
  • [26] Kagan, A. M. and Konikov, M. (2006). The structure of the UMVUEs from categorical Data., Theor. Probab. Appl. 50, 466–473.
  • [27] Kagan, A. M., Yu, T., Barron, A. and Madiman, M. (2012). Contribution to the theory of Pitman estimators., Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI) 408, 245–267.
  • [28] Kariya, T. (1989). Equivariant estimation in a model with an ancillary statistic., Ann. Statist. 17, 920–928.
  • [29] Klebanov, L. B. (1973). On the characterization of a family of distributions by the property of independence of statistics., Theor. Probab. Appl. 18, 608–611.
  • [30] Khan, R. A. (1968). A note on estimating the mean of a normal distribution with known coefficient of variation., J. Amer. Statist. Assoc. 63, 1039–1041.
  • [31] Lehmann, E. L. and Scheffé (1950). Completness, similar regions, and unbisased estimation: part I., Sankhya 10, 305–340.
  • [32] Lehmann, E. L. and Casella, G. (1998)., Theory of Point Estimation, 2nd ed. Springer, New York.
  • [33] Lehmann, E. L. and Romano, J. P. (2005)., Testing Statistical Hypotheses, 3rd ed. Springer, New York
  • [34] Linnik, Yu. V. (1968)., Statistical Problems with Nuisance Parameters. American Mathematical Society, Providence.
  • [35] Nayak, T. K. and Sinha, B. (2012). Some aspects of minimum variance unbiased estimation in presence of ancillary statistics., Stat. Prob. Letters 82, 1129–1135.
  • [36] Pitman, E. J. G. (1939a). The estimation of the location and scale parameters of a continuous population of any given form., Biometrika 30, 391–421.
  • [37] Pitman, E. J. G. (1939b). Tests of hypotheses concerning location and scale parameters., Biometrika 31, 200–215.
  • [38] Rao, C. R. (1952). Some theorems on minimum variance estimation., Sankhya Ser. A 12, 27–42.
  • [39] Unni, K. (1978). The theory of estimation in algebraic and analytical exponential families with applications to variance components models. PhD. Thesis, Indian Statistical Institute. Calcutta, India.
  • [40] Wijsman, R. A. (1958). Incomplete sufficient statistics and similar tests., Ann. Math. Statist. 29, 1028–1045.
  • [41] Zacks, S. (1971)., The Theory of Statistical Inference. Wiley, New York.