search

drizopoulos / GLMMadaptive

GLMMs with adaptive Gaussian quadrature
https://drizopoulos.github.io/GLMMadaptive/
60 stars 14 forks source link

A binomial model with log link function doesn't display summary #53

Closed scherbakovdmitri closed 7 months ago

scherbakovdmitri commented 7 months ago

Hello all, thanks for helpful package, glmer didn't converge, I am glad I found yours. But I have a problem, here is my model

`m2=GLMMadaptive::mixed_model(fixed=screen.completed~Variable ,control = list(#max_coef_value=1000,iter_EM=1000,iter_qN_outer=1000, verbose=T,

iter_qN=100,

                                     update_GH_every=40
                                     ),
                      random=~1|Department,
                      na.action=na.fail,
                      data=data
                      ,family=binomial(link='log')

)`

when I view summary I get error, so how do I plot confidence intervals for coefficients? Looks like the problem is only with log link function.

summary(m2) Error in solve.default(object$Hessian) : Lapack routine dgesv: system is exactly singular: U[1,1] = 0

drizopoulos commented 7 months ago

The model has not successfully converged.