cov2pcorr: Convert Covariance into Partial Correlation Matrix

In maotai: Tools for Matrix Algebra, Optimization and Inference

Description

Given a covariance matrix, return a partial correlation matrix that has unit diagonals. We strictly impose (and check) whether the given input is a symmetric matrix of full-rank.

Usage

cov2pcorr(mat)

Arguments

mat	a (p\times p) covariance matrix.

Value

a (p\times p) partial correlation matrix.

Examples

# generate an empirical covariance scaled
prep_mat = stats::cov(matrix(rnorm(100*10),ncol=10))
prep_vec = diag(as.vector(stats::runif(10, max=5)))
prep_cov = prep_vec%*%prep_mat%*%prep_vec

# compute correlation and partial correlation matrices
prep_cor = cov2corr(prep_cov)
prep_par = cov2pcorr(prep_cov)

# visualize
opar <- par(no.readonly=TRUE)
par(mfrow=c(1,3), pty="s")
image(prep_cov, axes=FALSE, main="covariance")
image(prep_cor, axes=FALSE, main="correlation")
image(prep_par, axes=FALSE, main="partial correlation")
par(opar)

maotai documentation built on Feb. 3, 2022, 5:09 p.m.