论文信息 - Eigenanalysis of symmetrizable matrix products: a result with statistical applications

Eigenanalysis of symmetrizable matrix products: a result with statistical applications

A theorem is proven which relates the matrices of eigenvalues and eigenvectors of matrix products AB7', ArB, BA', B'A if they are symmetrizable, that is if each product itself is expressible as the product of a symmetric and of a positive definite matrix. The result is used to derive properties of a number of different multivariate statistical techniques. Ke-,w,ords: canonical correlations, correspondence analysis. derived responses. eigenvalues, eigenvectors, principal components

Nanny Wermuth | N. Wermuth | H. Rüssmann | H. Rüssmann

[1] N. Wermuth,et al. Derived variables calculated from similar joint responses: some characteristics and examples , 1995 .

[2] N. Wermuth,et al. On the calculation of derived variables in the analysis of multivariate responses , 1992 .

[3] Christopher C. Paige,et al. The general linear model and the generalized singular value decomposition , 1985 .

[4] J. Gower. Some distance properties of latent root and vector methods used in multivariate analysis , 1966 .

[5] N. L. Johnson,et al. Linear Statistical Inference and Its Applications , 1966 .

[6] H. Schwerdtfeger. Direct Proof of Lanczos' Decomposition Theorem , 1960 .

[7] C. Lanczos,et al. Linear Systems in Self-Ad Joint Form , 1958 .

[8] H. Hotelling. Relations Between Two Sets of Variates , 1936 .

[9] C. Eckart,et al. The approximation of one matrix by another of lower rank , 1936 .

[10] H. Hirschfeld. A Connection between Correlation and Contingency , 1935, Mathematical Proceedings of the Cambridge Philosophical Society.

[11] Tosio Kato. Perturbation theory for linear operators , 1966 .

[12] H. Hotelling. Analysis of a complex of statistical variables into principal components. , 1933 .