simcor: Simulates a block correlation matrix

In MarkJBrewer/ICsims: Simulation framework allowing for between-sample heterogeneity in GLMs

Description

simcor generates a correlation matrix in the manner of Hardin et al (2013).

Usage

simcor(k = 6, size = c(10, 5, 8, 2, 15, 50), rho = c(0.7, 0.7, 0.5, 0.9,
  0.85, 0.4), delta = 0.39, epsilon = 0.99 - max(rho), eidim = 2)

Arguments

k	Integer. Number of block diagonals.
size	An integer vector of length k containing the number of rows (also columns) of a (square) block diagonal element of the final matrix, which will then be a square matrix of row length sum(size).
rho	A numeric vector of length k, containing the mean (off-diagonal) correlations in each block.
delta	Numeric. The mean correlation in the remaining elements of the final matrix, outside the block diagonals.
epsilon	Numeric. The degree of variability in the matrix.
eidim	The shape of the distribution of variability; if eidim==1, the variability around the mean is exactly epsilon but in random directions. Increasing values of eidim tend to centre the variation.

References

For full details, see Hardin, J., Garcia, S. R., and Golan, D. (2013). A method for generating realistic correlation matrices. Annals of Applied Statistics, 7(3):1733-1762. This function is not intended for use by the user.

MarkJBrewer/ICsims documentation built on May 7, 2019, 3:34 p.m.