NetFVar: Fréchet Variance for Networks

In frechet: Statistical Analysis for Random Objects and Non-Euclidean Data

Fréchet Variance for Networks

Description

Obtain Fréchet variance for graph Laplacian matrices, covariance matrices, or correlation matrices with respect to the Frobenius distance.

Usage

NetFVar(Ly = NULL)

Arguments

Ly	A list (length n) of m by m matrices or a m by m by n array where Ly[, , i] contains an m by m matrix, which can be either graph Laplacian matrices or covariance matrices or correlation matrices.

Value

A list containing the following fields:

NetFVar	A scalar holding the Fréchet variance.
NetFMean	A matrix holding the Fréchet mean.

Examples

set.seed(1)
n <- 100
U <- pracma::randortho(10)
Ly <- lapply(1:n, function(i) {
  U %*% diag(rexp(10, (1:10)/2)) %*% t(U)
})
res <- NetFVar(Ly)
res$NetFVar

frechet documentation built on May 29, 2024, 7:35 a.m.