rmatrix.spd_LogCholesky: Generate a set of random matrices on a matrix manifold

In linulysses/matrix-manifold: Basic Operations and Functions on Matrix Manifolds

Description

Generate a set of random matrices on a matrix manifold

Usage

## S3 method for class 'spd_LogCholesky'
rmatrix(mfd, n = 1, mu = NULL, sig = 1, drop = T)

Arguments

mfd	an object created by matrix.manifold
n	sample size
mu	the Frechet mean. If NULL is given, then it is the identity element, e.g., for SPD, it is the identity matrix
sig	the standard deviation of the normal distribution
drop	whether return the result as a matrix when n=1

Details

The generated samples have Frechet mean mu. The logarithmic maps of these samples at mu follow a isotropic D-dimensional normal distribution with isotropic variance sig, where D is the intrinsic dimension of the matrix manifold

Value

an M*N*n array of n matrices, where M*N is the dimensions of matrices

linulysses/matrix-manifold documentation built on Jan. 19, 2022, 3:56 p.m.