R: Generalized Linear Mixed Model Estimation via Monte Carlo EM
mcemGLM-package
R Documentation
Generalized Linear Mixed Model Estimation via Monte Carlo EM
Description
mcemGLM performs maximum likelihood estimation for logistic,
Poisson, and negative binomial regression when random effects are
present. The package uses an MCEM algorithm to estimate the model's
fixed paramters and variance components with their respective
standard errors.
A Wald test based anova is available to test significance of
multi-leveled variables and for multiple contrast testing.
Details
Package:
mcemGLM
Type:
Package
Version:
1.0
Date:
2015-05-22
License:
GPL (>= 2)
Author(s)
Felipe Acosta Archila
Maintainer: Felipe Acosta Archila <acosta@umn.edu>
Examples
set.seed(123)
x <- rnorm(30, 10, 1)
z <- factor(rep(1:6, each = 5))
obs <- sample(0:1, 30, TRUE)
fit <- mcemGLMM(obs ~ x, random = ~ 0 + z, family = "bernoulli",
vcDist = "normal", controlEM = list(EMit = 15, MCit = 10000),
initial = c(3.30, -0.35, 0.005))
summary(fit)
anova(fit)