plsform: Create the structure of a linear mixed model from...

In lme4/lme4pureR: lme4 in pure R

Description

A pure R implementation of the penalized least squares (PLS) approach to evaluation of the deviance or the REML criterion for linear mixed-effects models.

Usage

plsform(formula, data, REML = TRUE, weights, offset, sparseX = FALSE,
  family = gaussian, ...)

Arguments

formula	a two-sided model formula with random-effects terms and, optionally, fixed-effects terms.
data	a data frame in which to evaluate the variables from form
REML	calculate REML deviance?
weights	prior weights
offset	offset
sparseX	should X, the model matrix for the fixed-effects coefficients be sparse?
...	additional arguments

Value

a list with:

• X Fixed effects model matrix

• y Observed response vector

• fr Model frame

• call Matched call

• REML Logical indicating REML or not

• weights Prior weights or NULL

• offset Prior offset term or NULL

• Zt Transposed random effects model matrix

• Lambdat Transposed relative covariance factor

• theta Vector of covariance parameters

• lower Vector of lower bounds for theta

• upper Vector of upper bounds for theta

• thfun A function that maps theta into the structural non-zero elements of Lambdat, which are stored in slot(Lambdat, 'x')

Examples

form <- Reaction ~ Days + (Days|Subject)
data(sleepstudy, package="lme4")
ll <- plsform(form, sleepstudy, REML=FALSE)
names(ll)

lme4/lme4pureR documentation built on May 21, 2019, 7:34 a.m.