FG: Joint Diagonalization of Real Positive-definite Matrices

In JADE: Blind Source Separation Methods Based on Joint Diagonalization and Some BSS Performance Criteria

Joint Diagonalization of Real Positive-definite Matrices

Description

This is a slightly modified version of Flury's FG algorithm for the joint diagonalization of k positive-definite matrices. The underlying function is written in C.

Usage

FG(X, weight = NULL, init = NULL, maxiter = 100, eps = 1e-06, na.action = na.fail)

Arguments

X	A matrix of k stacked pxp matrices with dimension c(kp,p) or an array with dimension c(p,p,k).
weight	A vector of length k to give weight to the different matrices, if NULL, all matrices have equal weight.
init	Initial value for the orthogonal matrix to be estimated, if NULL, the identity matrix is used.
maxiter	Maximum number of iterations.
eps	Convergence tolerance.
na.action	A function which indicates what should happen when the data contain 'NA's. Default is to fail.

Value

A list with the components

V	An orthogonal matrix.
D	A stacked matrix with the diagonal matrices or an array with the diagonal matrices. The form of the output depends on the form of the input.
iter	The Fortran function returns also the number of iterations.

Author(s)

Jari Miettinen

References

Flury, B. D. (1998), Common principal components and related models, Wiley, New York.

See Also

rjd, rjd.fortran

JADE documentation built on Sept. 18, 2023, 1:06 a.m.