rcov: Generate a random covariance matrix

In iamamutt/mejr: General Purpose R Functions

Description

Generate a random covariance matrix

Usage

rcov(n, size, regularization = 1, concentration = 1, tau_shape = 1,
  tau_scale = 1, trace = size)

Arguments

n	Number of random matrices to generate
size	Number of columns/variances in a covariance matrix.
regularization	Positive scalar. Controls correlation strength. 1 is uniform over correlation matrices. Greater than 1 has weaker correlations.
concentration	Positive scalar. Controls the spread of proportions of the total variance. 1 is uniform, less than 1 creates heterogeneity, greater than 1 makes variances homogeneous.
tau_shape	Positive scalar. Controls total variance. Shape of a gamma distribution, defaults to exponential. Influences dispersion around mean variance.
tau_scale	Positive scalar. Controls total variance. Scale of a gamma distribution, defaults to exponential. sqrt(sum(diagonal)) if tau_shape=1 and trace=1.
trace	Positive scalar. Total variance or the sum of variances on the diagonal

Value

A covariance matrix or a three-dimensional array of matrices if n>1

Examples

rcov(n=5, size=4, trace=10)

iamamutt/mejr documentation built on May 18, 2019, 1:27 a.m.