sigma.lmm: Extract The Residuals Variance-Covariance Matrix From a...

In LMMstar: Repeated Measurement Models for Discrete Times

Extract The Residuals Variance-Covariance Matrix From a Linear Mixed Model

Description

Extract the unique set of residuals variance-covariance matrices or the one relative to specific clusters.

Usage

## S3 method for class 'lmm'
sigma(
  object,
  cluster = NULL,
  p = NULL,
  chol = FALSE,
  inverse = FALSE,
  simplify = TRUE,
  ...
)

Arguments

object	a lmm object.
cluster	[character, data.frame, NULL] identifier of the cluster(s) for which to extract the residual variance-covariance matrix. For new clusters, a dataset containing the information (cluster, time, strata, ...) to be used to generate the residual variance-covariance matrices. When NULL, will output complete data covariance patterns.
p	[numeric vector] value of the model coefficients at which to evaluate the residual variance-covariance matrix. Only relevant if differs from the fitted values.
chol	[logical] Output the cholesky factorization of the variance-covariance matrix.
inverse	[logical] Output the matrix inverse of the variance-covariance matrix.
simplify	[logical] When there is only one variance-covariance matrix, output a matrix instead of a list of matrices.
...	Not used. For compatibility with the generic method.

Value

A list where each element contains a residual variance-covariance matrix. Can also be directly a matrix when argument is simplify=TRUE and there is a single residual variance-covariance matrix.

Examples

## simulate data in the long format
set.seed(10)
dL <- sampleRem(100, n.times = 3, format = "long")
dL$id.fac <- paste0("id",dL$id)

## fit Linear Mixed Model
eUN.lmm <- lmm(Y ~ X1 + X2 + X5, repetition = ~visit|id.fac,
               structure = "UN", data = dL, df = FALSE)

## extract residuals variance covariance matrix
sigma(eUN.lmm) ## unique patterns
sigma(eUN.lmm, cluster = c("id1","id5")) ## existing clusters
sigma(eUN.lmm, cluster = dL[1:7,,drop=FALSE]) ## new clusters

LMMstar documentation built on Nov. 9, 2023, 1:06 a.m.