Skip to main content
Log in

Craig-Sakamoto's theorem for the Wishart distributions on symmetric cones

  • Published:
Annals of the Institute of Statistical Mathematics Aims and scope Submit manuscript

Abstract

A version of Craig-Sakamoto's theorem says essentially that ifX is aN(O,I n ) Gaussian random variable in ∝n, and ifA andB are (n, n) symmetric matrices, thenX′AX andX′BX (or traces ofAXX′ andBXX′) are independent random variables if and only ifAB=0. As observed in 1951, by Ogasawara and Takahashi, this result can be extended to the case whereXX′ is replaced by a Wishart random variable. Many properties of the ordinary Wishart distributions have recently been extended to the Wishart distributions on the symmetric cone generated by a Euclidean Jordan algebraE. Similarly, we generalize there the version of Craig's theorem given by Ogasawara and Takahashi. We prove that ifa andb are inE and ifW is Wishart distributed, then Tracea.W and Traceb.W are independent if and only ifa.b=0 anda.(b.x)=b.(a.x) for allx inE, where the. indicates Jordan product.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  • Casalis, M. (1990). Familles exponentielles naturelles invariantes par un groupe, Ph.D. Thesis, Laboratoire de Statistique et Probabilités, Université Paul Sabatier, Toulouse.

    Google Scholar 

  • Casalis, M. (1991). Les familles exponentielles à variance quadratique homogène sont des lois de Wishart sur un cöne symétrique.C. R. Acad. Sc. Paris,312, 537–540.

    Google Scholar 

  • Casalis, M. and Letac, G. (1994). Characterization of the Jorgensen set in the generalized linear model,Test,3, 145–161.

    Google Scholar 

  • Driscoll, M. F. and Gundberg, W. R. (1986). A history of the development of Craig's theorem,Amer. Statist.,40, 65–70.

    Google Scholar 

  • Faraut, J. (1988). Algèbres de Jordan et cônes symétriques, Notes d'un Cours d'Eté du Centre International de Mathématiques Pures et Appliquées, Département de Mathématiques, Université de Poitiers.

  • Faraut, J. and Koranyi, A. (1994).Analysis on Symmetric Cones, Oxford University Press, Oxford.

    Google Scholar 

  • Gindikin, S. (1975). Invariant generalized functions in homogeneous spaces.J. Funct. Anal. Appl.,9, 50–52.

    Google Scholar 

  • Jacobson, N. (1968). Structure and representations of Jordan algebras,Amer. Math. Soc. Collog. Publ.,XXXXIX, Providence, Rhode Island.

  • Jensen, S. (1988). Covariance hypotheses which are linear in both the covariance and in the inverse covariance,Ann. Statist.,16, 302–322.

    Google Scholar 

  • Kallenberg, O. (1991). On an independence criterion for multiple Wiener integrals,Ann. Probab.,19, 83–85.

    Google Scholar 

  • Lancaster, H. O. (1954). Traces and cumulants of quadratic forms in the normal variables,J. Roy. Statist. Soc. Ser. B,16, 247–254.

    Google Scholar 

  • Letac, G. (1994). Les familles exponentielles statistiques invariantes par les groupes du cône et du paraboloïde de révolution,Volume in Honor of L. Takaćs (eds. J. Galambos and J. Gani), Applied Probability Trust, Sheffield.

    Google Scholar 

  • Lévy, P. (1948). The arithmetical character of the Wishart distribution,Proc. Cambridge Philos. Soc.,44, 255–297.

    Google Scholar 

  • Massam, H. (1994). An exact decomposition theorem and a unified view of some related distributions for a class of exponential transformation models on symmetric cones,Ann. Statist.,22, 369–394.

    Google Scholar 

  • Massam, H. and Neher, E. (1994). On transformations and determinants of Wishart variables on symmetric cones (submitted).

  • Matusita, K. (1949). Note on the independence of certain statistics,Ann. Inst. Statist. Math.,1, 79–82.

    Google Scholar 

  • Muirhead, R. J. (1982).Aspects of Multivariate Statistical Theory, Wiley, New York.

    Google Scholar 

  • Ogasawara, T. and Takahashi, M. (1951). Independence of quadratic quantities in a normal system,Journal of Science of the Hiroshima University, Series A,15, 1–9.

    Google Scholar 

  • Ogawa, J. (1949). On the independence of bilinear and quadratic forms of a random sample from a normal population,Ann. Inst. Statist. Math.,1, 83–108.

    Google Scholar 

  • Ogawa, J. (1993). A history of the development of Craig-Sakamoto's theorem viewed from Japanese standpoint,Proc. Inst. Statist. Math.,21, 47–59 (in Japanese).

    MathSciNet  Google Scholar 

  • Olkin, I. (1990). Interface between statistics and linear algebra,Proc. Sympos. Appl. Math.,40, 233–256.

    Google Scholar 

  • Rothaus, O. (1968). Some properties of Laplace transforms of measures,Trans. Amer. Math. Soc.,131, 163–169.

    Google Scholar 

  • Sakamoto, H. (1944). On the independence of statistics, Research Memo.1, The Institute of Statistical Mathematics, Tokyo (in Japanese).

    Google Scholar 

  • Satake, I. (1980).Algebraic Structures on Symmetric Domains, Iwanami-Shoten, Tokyo and Princeton University Press, Princeton.

    Google Scholar 

  • Shanbhag, D. N. (1988). The Davidson-Kendall problem and related results on the structure of the Wishart distribution,Austral. J. Statist.,30A, 272–280.

    Google Scholar 

  • Ustünel, A. S. and Zakaï, M. (1989). On independence and conditioning on Wiener space,Ann. Probab.,17, 1441–1453.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Additional information

Partially supported by NATO grant 92.13.47.

About this article

Cite this article

Letac, G., Massam, H. Craig-Sakamoto's theorem for the Wishart distributions on symmetric cones. Ann Inst Stat Math 47, 785–799 (1995). https://doi.org/10.1007/BF01856547

Download citation

  • Received:

  • Revised:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF01856547

Key words and phrases

Navigation