Dear Statalisters,
I search a program to estimate the marginal likelihood of a mixt logistic
model with 2 random variables (following a gaussian distribution). The two
random variables are correlated so the estimation of this marginal
likelihood can be realized by two-dimensional gauss Hermite quadratures (or
other method ?).
Tha aim of my work is to find an alternative to GLLAMM to estimate fastly
the parameters of a such model. GLLAMM is not a solution for me, because
this program estimate the marginal likelihood at each step to estimate the
parameters of the model, and me, I search a program to estimate this
marginal likelihood conditionaly to known parameters (estimated a priori by
GEE).
