On multi-set canonical correlation analysis

Two- and multi-set canonical correlation analysis (CCA) and (MCCA) techniques are used to find linear combinations that give maximal multivariate differences. This paper describes methods for deriving MCCA dynamical systems which converge to the desired canonical variates and canonical correlations. Unconstrained and constrained optimization methods over quadratic constraints are applied to derive several dynamical systems that converge to a solution of a generalized eigenvalue problem. These include merit functions that are based on generalized Rayleigh quotient, and logarithmic generalized Rayleigh quotient.

[1]  N. L. Johnson,et al.  Multivariate Analysis , 1958, Nature.

[2]  Mohammed A. Hasan.  A new approach for computing canonical correlations and coordinates , 2004, 2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512).

[3]  Jan de Leeuw,et al.  Non-linear canonical correlation , 1983 .

[4]  Anja Vogler,et al.  An Introduction to Multivariate Statistical Analysis , 2004 .

[5]  Jan de Leeuw,et al.  Use of the multinomial jack-knife and bootstrap in generalized non-linear canonical correlation analysis , 1988 .

[6]  Winson Taam,et al.  Non-linear canonical correlation analysis with a simulated annealing solution , 1992 .

[7]  H. Hotelling.  The most predictable criterion. , 1935 .

[8]  M.A. Hasan,et al.  Gradient dynamical systems for principal singular subspace analysis , 2008, 2008 American Control Conference.

[9]  Weiping Li,et al.  Applied Nonlinear Control , 1991 .

[10]  Colin Fyfe,et al.  A neural implementation of canonical correlation analysis , 1999, Neural Networks.

[11]  J. Magnus,et al.  Matrix Differential Calculus with Applications in Statistics and Econometrics , 1991 .

[12]  Chong-sun Kim.  Canonical Analysis of Several Sets of Variables , 1973 .

[13]  Colin Fyfe,et al.  Kernel and Nonlinear Canonical Correlation Analysis , 2000, IJCNN.

[14]  Colin Fyfe,et al.  A canonical correlation neural network for multicollinearity and functional data , 2004, Neural Networks.

[15]  Allan Aasbjerg Nielsen,et al.  Multiset canonical correlations analysis and multispectral, truly multitemporal remote sensing data , 2002, IEEE Trans. Image Process..

[16]  Mohammed A. Hasan,et al.  Unconstrained functional criteria for canonical correlation analysis , 2005, 2005 IEEE International Symposium on Circuits and Systems.

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

[18]  M. Hasan,et al.  Dynamical Systems for Principal Singular Subspace Analysis , 2006, Fourth IEEE Workshop on Sensor Array and Multichannel Processing, 2006..