Biometrika Advance Access published online on November 11, 2008
Biometrika, doi:10.1093/biomet/asn052
Article |
Covariance reducing models: An alternative to spectral modelling of covariance matrices
School of Statistics, University of Minnesota, Minneapolis, Minnesota 55455, U.S.A. dennis{at}stat.umn.edu
Facultad de Ingeniería Química, Universidad Nacional del Litoral and Instituto Matemática Aplicada Litoral, CONICET, Santa Fe, Argentina liliana.forzani{at}gmail.com
Received for publication 1 December 2007. Revision received 1 April 2008.
We introduce covariance reducing models for studying the sample covariance matrices of a random vector observed in different populations. The models are based on reducing the sample covariance matrices to an informational core that is sufficient to characterize the variance heterogeneity among the populations. They possess useful equivariance properties and provide a clear alternative to spectral models for covariance matrices.
Key Words: Central subspace • Dimension reduction • Envelopes • Grassmann manifolds • Reducing subspaces
References
-
Boik R. J. Spectral models for covariance matrices. Biometrika (2002) 89:159–82.
Burnham K., Anderson D. Model Selection and Multimodel Inference. (2002) New York: Wiley.
Chikuse Y. Statistics on Special Manifolds (2003) New York: Springer.
Conway J. B. A Course in Functional Analysis (1990) 2nd ed. New York: Springer.
Cook R. D. Using dimension-reduction subspaces to identify important inputs in models of physical systems. Proceedings of the section on Physical and Engineering Sciences (1994) Alexandria, VA: American Statistical Association. 18–25.
Cook R. D. Regression Graphics. (1998) New York: Wiley.
Cook R.D., Li B., Chiaromonte F. Dimension reduction without matrix inversion. Biometrika (2007) 94:569–84.
Eaton M. Multivariate Statistics (1983) New York: Wiley.
Flury B. Some relations between the comparison of covariance matrices and principal component analysis. Comp. Statist. Data Anal. (1983) 1:97–109.[CrossRef]
Flury B. Common principal components in k groups. J. Am. Statist. Assoc. (1984) 79:892–98.[CrossRef][Web of Science]
Flury B. Two generalizations of the common principal component model. Biometrika (1987) 74:59–69.
Flury B. Common Principal Components and Related Multivariate Models. (1988) New York: Wiley.
Greub W. Linear Algebra. (1981) New York: Springer.
Houle D., Mezey J., Galpern P. Interpretation of the results of common principal component analysis. Evolution (2002) 56:433–40.[CrossRef][Web of Science][Medline]
Jensen S. T., Madsen J. Estimation of proportional covariance matrices. Ann. Statist. (2004) 32:219–32.[CrossRef]
Liu X., Srivastava A., Gallivan K. Optimal linear representations of images for object recognition. IEEE Trans. Pat. Anal. Mach. Intel. (2004) 26:662–66.[CrossRef]
Mezey J. G., Houle D. Comparing G matrices: Are common principal components informative? Genetics (2003) 165:411–25.
Muirhead R. J. Aspects of Multivariate Statistical Theory (1982) New York: Wiley.
Phillips P. C., Arnold S. J. Hierarchical comparison of genetic variance–covariance matrices I. Using the Flury hierarchy. Evolution (1999) 53:1506–15.[CrossRef][Web of Science]
Rao C. R. Linear Statistical Inference and its Applications (1973) 2nd ed. New York: Wiley.
Schott J. R. Some tests for common principal component subspaces. Biometrika (1991) 75:229–36.[CrossRef]
Schott J. R. Partial common principal component subspaces. Biometrika (1999) 86:899–908.
Schott J. R. Weighted chi-squared test for partial common principal component subspaces. Biometrika (2003) 90:411–21.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||