Hostname: page-component-84b7d79bbc-7nlkj Total loading time: 0 Render date: 2024-07-29T13:00:23.195Z Has data issue: false hasContentIssue false
  • English
  • Français

Improved Berry-Esseen-Chebyshev Bounds with Statisical Applications

Published online by Cambridge University Press:  18 October 2010

Jean-Marie Dufour  and
Marc Hallin
Jean-Marie Dufour
Affiliation:
Université de Montréal
Marc Hallin
Affiliation:
Université Libre de Bruxelles
Get access Rights & Permissions [Opens in a new window]

Abstract

A Sharpening Of Nonuniform bounds of the Berry-Esseen type initially obtained by Esseen and later generalized by Kolodjažnyĭ–who also proved that they are, in some sense, optimal–is proposed. Further, the corresponding inequalities are shown to provide uniformly improved Chebyshev bounds for the tail probabilities of the distribution functions to be approximated. In contrast with most results on Berry–Esseen bounds, which emphasize rates of convergence to normality, the bounds proposed are sufficiently explicit to allow the computation of numerical bounds on a distribution function. For example, they can be applied to the sum of a small number of independent random variables. The bounds are easy to compute and can be used in confidence estimation as well as in testing problems. Applications include signed-rank tests, permutation tests, and the chi-square approximation to Bartlett's test statistic for the homogeneity of several variances.

Type
Articles
Information
Econometric Theory , Volume 8 , Issue 2 , June 1992 , pp. 223 - 240
Copyright
Copyright © Cambridge University Press 1992

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

REFERENCES

1.Bartlett, M.S.Properties of sufficiency and statistical tests. Proceedings of the Royal Society of London A 160 (1937): 268–282.Google Scholar
2.Bhattacharya, R.N. & Rao, R.R.. Normal Approximation and Asymptotic Expansions. New York: Wiley, 1976.Google Scholar
3.Bickel, P. & Freedman, D.A.. Some asymptotic theory for the bootstrap. The Annals of Statistics 9 (1981): 1196–1217.CrossRefGoogle Scholar
4.Bikelis, A.On an estimate of the remainder term in the central limit theorem.Litovskii Matematiceskii Shornik 4 (1964): 303–308. (In Russian)Google Scholar
5.Bikelis, A.Estimates of the remainder term in the central limit theorem. Litovskit Mate-maticeskii Shornik 6 (1966): 323–346. (In Russian).Google Scholar
6.Bolthausen, E.An estimate of the remainder in a combinatorial central limit theorem. Zeitschrift fur Wahrscheinlichkeitstheorie und verwandte Gebiete 66 (1984): 379–386.CrossRefGoogle Scholar
7.Callaert, H. & Janssen, P.. The Berry-Esseen theorem for U-statistics. The Annals of'Statistics 6 (1978): 417–421.CrossRefGoogle Scholar
8.Chan, Y.-K. & Wierman, J.. On the Berry-Esseen theorem for U-statistics. The Annals of Probability 5 (1977): 136–139.CrossRefGoogle Scholar
9.Diciccio, T.J. & Romano, J.P.. A review of bootstrap confidence intervals. Journal of the Royal Statistical Society, Series B 50 (1988): 338–354.Google Scholar
10.Dufour, J.M. & Hallin, M.. Simple exact bounds for distributions of linear signed rank statistics. To appear in Journal of Statistical Planning and Inference (1992).CrossRefGoogle Scholar
11.Efron, B.Nonparametric standard errors and confidence intervals (with discussion). The Canadian Journal of Statistics 9 (1981): 139–172.CrossRefGoogle Scholar
12.Esseen, C.G.Fourier analysis of distribution functions. A mathematical study of the La-place-Gaussian law. Acta Mathematica 11 (1945): 1–125.CrossRefGoogle Scholar
13.Friedrich, K.O.A Berry-Esseen bound for functions of independent random variables. The Annals of Statistics 17 (1989): 170–183.CrossRefGoogle Scholar
14.Hall, P.Two-sided bounds for nonuniform rates of convergence in the central limit theorem. Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete 65 (1983): 61–72.CrossRefGoogle Scholar
15.Hall, P.Theoretical comparison of bootstrap confidence intervals (with discussion). The Annals of Statistics 16 (1988): 927–985.Google Scholar
16.Helmers, R.A Berry-Esseen theorem for linear combinations of order statistics. The Annals of Probability 9 (1981): 342–347.CrossRefGoogle Scholar
17.Helmers, R. & Huskova, M.. A Berry-Esseen bound for L-statistics with unbounded weight functions. In Mandl, P. and Huskova, M. (eds.), Asymptotic Statistics 2, Proceedings of the Third Prague Symposium on Asymptotic Statistics, pp. 93–101. Amsterdam: North-Holland, 1984.Google Scholar
18.Heimers, R. & Zwet, W.R. van. The Berry-Esseen bound for U-statistics. In Gupta, S.S. and Berger, J.O. (eds.), Statistical Decision Theory and Related Topics III, pp. 497–512. New York: Academic Press, 1982.Google Scholar
19.Heyde, C.C.A nonunifonn bound on convergence to normality. The Annals of Probability 3(1975): 903–907.CrossRefGoogle Scholar
20.Kolodjažnyï, S.F.A generalization of a theorem of Esseen. Vestnik Leningradskogo Universiteta. Matematika 23, 13 (1968): 28–33. Vestnik Leningrad University, Mathematics 1 (1974): 189–195.Google Scholar
21.Korolyuk, V.S. & Borovskikh, Yu. V.. Approximation of nondegeeerate U-statistics. Theory of Probability and its Applications 30 (1985): 439–450.CrossRefGoogle Scholar
22.Manoukian, E.B.Bound on the accuracy of Bartlett's chi-square approximation for testing the homogeneity of variances. SIAM Journal of Applied Mathematics 42 (1982): 575–587.CrossRefGoogle Scholar
23.Michel, R.Nonuniform central limit bounds with applications to probabilities of deviations.The Annals of Probability 4 (1976): 102–106.CrossRefGoogle Scholar
24.Michel, R.On the accuracy of nonuniform Gaussian approximation to the distribution functions of sums of independent and identically distributed random variables. Zeitschrift fur Wahrscheinlichkeitstheorie und verwandte Gebiete 35 (1976): 337–347.CrossRefGoogle Scholar
25.Michel, R.On the constant in the nonuniform version of the Berry-Esseen theorem. Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete 55 (1981): 109–117.CrossRefGoogle Scholar
26.Nagaev, A.V.Some limit theorems for large deviations. Theory of Probability and its Applications 10 (1965): 214–235.CrossRefGoogle Scholar
27.Osipov, L.V.Asymptotic expansions in the central limit theorem. Vestnik Leningradskogo Universiteta. Matematika, Mehanika, Astronomija 19, 4 (1967): 45–62. (In Russian)Google Scholar
28.Osipov, L.V. & Petrov, V.V.. On an estimate of the remainder term in the central limit theorem. Theory of Probability and Its Applications 12 (1967): 281–286.CrossRefGoogle Scholar
29.Paditz, L. Über die Annäherung der Verteilungsfunktionen von Summen unabhängiger Zufallsgrōssen gegen unbegrenzt teilbare Verteilungsfunktionen unter besonderer Beachtung der Verteilungsfunktion der standardisierten Normalverteilung. Dissertation A TU Dresden 1977.Google Scholar
30.Paditz, L.Abschätzungen der Konvergenzgeschwindigkeit zur Normalverteilung unter Voraussetzung linreitiger Momente. Mathematische Nachrichten 82 (1978): 131–156.CrossRefGoogle Scholar
31.Petrov, V.V.Sums of Independent Random Variables. Berlin and New York: Springer-Verlag, 1975.Google Scholar
32.Pipiras, V.On the remainder terms in the asymptotic expansion of the distribution function of the sum of independent random variables. Litovskiľ Matematiceskii Sbornik 10 (1970): 135–159. (In Russian)Google Scholar
33.Puri, M.L. & Ralescu, S.S.. On the degeneration of the variance in the asymptotic normality of signed rank statistics. In Kallianpur, G., Krishnaiah, P.R. and Ghosh, J.K. (eds.), Statistics and Probability: Essays in Honor of C.R. Rao, pp. 591–607. Amsterdam: North-Holland, 1982.Google Scholar
34.Puri, M.L. & Ralescu, S.S.. Centering of signed rank statistic with continuous score-generating function. Theory of Probability and its Applications 29 (1984): 580–584.Google Scholar
35.Puri, M.L. & Seoh, M.. Berry-Esseen theorems for signed linear rank statistics with regression constants. In Revesz, P. (ed.), Limit Theorems in Probability and Statistics, pp. 875–905. Amsterdam: North-Holland, 1984.Google Scholar
36.Puri, M.L. & Seoh, M.. Edgeworth expansions for signed linear rank statistics with regression constants. Journal of Statistical Planning and Inference 10 (1984): 137–149.CrossRefGoogle Scholar
37.Puri, M.L. & Seoh, M.. Edgeworth expansions for signed linear rank statistics under near location alternatives. Journal of Statistical Planning and Inference 10 (1984): 289–309.CrossRefGoogle Scholar
38.Puri, M.L.& Seoh, M.. On the rate of convergence to normality for generalized linear rank statistics. Annals of the Institute of Statistical Mathematics 37 (1985): 51–69.CrossRefGoogle Scholar
39.Puri, M.L. & Wu, T.-J.. The order of normal approximation for signed linear rank statistics. Theory of Probability and its Applications 31 (1986): 145–151.CrossRefGoogle Scholar
40.Ralescu, S. & Puri, M.L.. On the rate of convergence in the central limit theorem for signed rank statistics. Advances in Applied Mathematics 6 (1985): 23–51.CrossRefGoogle Scholar
41.Seoh, M.Berry-Esseen-type bounds for signed linear rank statistics with a broad range of scores. The Annals of Statistics 18 (1990): 1483–1490.CrossRefGoogle Scholar
42.Seoh, M. & Puri, M.L.. Berry-Esseen theorems for signed linear rank statistics under near location alternatives. Studia Scientiarum Mathematicarum Hungarica 20 (1985): 197–211.Google Scholar
43.Seoh, M., Ralescu, S.S. & Puri, M.L.. Cramer type large deviations for generalized rank statistics. The Annals of Probability 13 (1985): 115–125.CrossRefGoogle Scholar
44.van Beek, P. Fourier-analytische Methoden zur Verscharfung der Berry-Esseen Schranke. Doctoraf Dissertation, Bonn, 1971.Google Scholar
45.van Beek, P.An application of Fourier methods to the problem of sharpening Berry-Esseen inequality. Zeitschrift für Wahrscheinlichkeitstheorie und verwandte Gebiete 23 (1972): 187–196.CrossRefGoogle Scholar
46.van Zwet, W.R.A Berry-Esseen bound for symmetric statistics. Zeitschriftfur Wahrschein-lichkeitstheorie und verwandte Gebiete 66 (1984): 425–440.CrossRefGoogle Scholar
47.Wu, T.-J.A large deviation result for signed linear rank statistics under the symmetry hypothesis. The Annals of Statistics 14 (1986): 774–780.CrossRefGoogle Scholar
48.Wu, T.-J.An Lp error bound in normal approximation for signed linear rank statistics. Sankhya A 49 (1987): 122–127.Google Scholar
49.Zolotarev, V.M.A one-sided interpretation and refinements of certain Chebyshev-type inequalities. Litovskiľ Matematiceskii Sbornik 5 (1965): 233–250. Selected Translations in Mathematical Statistics and Probability 12 (1973): 25–45.Google Scholar