Abstract
The problem of testing multinormality against some alternatives invariant with respect to a subgroup of affine transformations is studied. Using the Laplace expansion for integrals, some approximations to the most powerful invariant (MPI) tests are derived. The cases of bivariate exponential and bivariate uniform alternatives are studied in detail, whereas higher-dimensional extensions are only outlined. Those alternatives are irregular and need a special treatment because of the dependence of their supports on the unknown parameters. It is shown that likelihood ratio (LR) tests are asymptotically equivalent to the MPI tests. A Monte Carlo simulation study shows that the powers of the LR tests are very close to the powers of the MPI tests, even in small samples.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Andersson S (1982) Distributions of maximal invariants using quotient measures. Ann Stat 10:955–961
Arnold BC, Balakrishnan N (1989) Relations, bounds and approximations for order statistics. Lecture notes in statistics, vol 53. Springer, Heidelberg
Barbe P (1993) Limiting distribution of the minimal spacing. Math Methods Stat 3:306–325
Barndorff-Nielsen OE, Cox DR (1989) Asymptotic techniques for use in statistics. Chapman and Hall, London
Barndorff-Nielsen OE, Cox DR (1994) Inference and asymptotics. Chapman and Hall, London
Barndorff-Nielsen OE, Jupp PE (1988) Differential geometry, profile likelihood, L-sufficiency and composite transformation models. Ann Stat 16:1009–1043
Barndorff-Nielsen OE, Blæsild P, Eriksen PS (1989) Decomposition and invariance of measures with a view to statistical transformation models. Lecture Notes in Statistics, vol 58. Springer, Heidelberg
Bleistein N, Handelsman RA (1975) Asymptotic expansions of integrals. Holt, Rinehart and Winston, New York
Ducharme GR, Frichot B (2003) Quasi-most powerful invariant goodness-of-fit tests. Scand J Stat 30:399–414
Eaton ML (1983) Multivariate statistics. A vector space approach. Wiley, New York
Fisher RA (1936) The use of multiple measurements in taxonomic problems. Ann Eugen 7/II:179–188
Henze N (2002) Invariant tests for multivariate normality: a critical review. Stat Pap 43:467–506
Mann HB, Wald A (1943) On stochastic limit and order relationship. Ann Math Stat 14:217–226
Mardia KV (1970) Measures of multivariate skewness and kurtosis with applications. Biometrika 57:519–530
Pace L, Salvan A (1997) Principles of statistical inference. World Scientific, Singapore
Pace L, Salvan A, Ventura L (2006) Likelihood-based discrimination between separate scale and regression models. J Stat Plan Inference 136:3539–3553
Press WH, Teukolsky SA, Vetterling WT, Flannery BP (1995) Numerical recipies in C. The art of scientific computing, 2nd edn. Cambridge University Press, New York
Serfling RJ (1980) Approximation theorems of mathematical statistics. Wiley, New York
Shun Z, McCullagh P (1995) Laplace approximation of high-dimensional integrals. J R Stat Soc B 57:749–760
Szkutnik Z (1987) On invariant tests for multidimensional normality. Probab Math Stat 8:1–10
Szkutnik Z (1988) Most powerful invariant tests for binormality. Ann Stat 16:292–301
Szkutnik Z (1992) Special capacities, the Hunt–Stein theorem and transformation groups. Ann Stat 20:1120–1128
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Majerski, P., Szkutnik, Z. Approximations to most powerful invariant tests for multinormality against some irregular alternatives. TEST 19, 113–130 (2010). https://doi.org/10.1007/s11749-008-0136-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11749-008-0136-4