cov_check: Covariance Check

In TheLondonPeacock/SeaClass: Interactive GUI for analyzing a binary response variable

Description

Checks a covariance matrix and identifies variables which should be dropped due to zero variance or high dependence.

Usage

cov_check(cov, cut_point = 0.9999)

Arguments

cov	a matrix.
cut_point	threshold which signifies two variables are highly dependent and one should be omitted. Specifically, when a group of variables are highly dependent, only the first variable is kept and all others are dropped. Defaults to 0.9999.

Details

Used to flag variables which will likely cause the covariance matrix to be singular or near singular; and hence, non-invertible.

Value

• remove_vars locations of variables to be dropped.

• remove_length number of variables to be dropped.

• remove_text a message detailing the recommended action.

See Also

cov_diag_check, cov_offdiag_check

Examples

cov1 <- matrix(c(1,0,0,0,0,0,0,0,0,0,1,1,0,0,1,1),ncol=4,nrow=4)
cov2 <- diag(c(1,1,1,1))
cov_check(cov1, cut_point = 0.9999)
cov_check(cov2, cut_point = 0.9999)

TheLondonPeacock/SeaClass documentation built on May 8, 2019, 3:42 p.m.