randogit: MCEM resolution for a probit maximum likelihood
Description
Based on Booth and Hobert (JRSS B, 1999), this function evaluates the maximum
likelihood estimate of a simulated probit model with random effects. The
random effects are simulated by a MCMC algorithm.
Usage
randogit(Tem = 10^3, Tmc = 10^2)
Arguments
Tem
starting number of MCEM iterations
Tmc
number of Monte Carlo points in the likelihood approximations
Value
The function returns two plots, one of $(beta,sigma)$ and one
of the true likelihood $L(beta,sigma,u0)$, where $u0$
is the true vector of random effects.
References
From Chapter 2 of EnteR Monte Carlo Statistical Methods