| booth {bernor} | R Documentation |
This data set contains a simulated data from the paper of Booth and Hobert referenced below.
data(booth)
A list containing components:
y | response matrix |
x | fixed effect model matrix |
z | random effect model matrix |
i | random effect index vector |
mu0 | “ simulation truth” fixed effect parameter |
sigma0 | “ simulation truth” random effect parameter |
theta0 | “ simulation truth” parameter vectorc(mu0, sigma0) |
theta.hat.exact | The maximum likelihood parameter vector calculated using the exact likelihood (calculated by numerical integration, not Monte Carlo) |
info0 | the expected Fisher information matrix for sample
size one (one column of y) at the
“simulation truth” parameter vector |
bigw0 | the expected “ big W” matrix for sample size one at the “simulation truth” parameter vector |
info.hat.exact | the expected Fisher information matrix for
sample size one (one column of y) at the
theta.hat.exact parameter vector |
bigw.hat.exact | the expected “ big W” matrix for sample size one at thetheta.hat.exact parameter vector
|
The structure of these objects is described in the documentation
for bnlogl and bnbigw.
Booth, J. G. and Hobert, J. P. (1999). Maximizing generalized linear mixed model likelihoods with an automated Monte Carlo EM algorithm. Journal of the Royal Statistical Society Series B (Statistical Methodology) 61, 265–285.