Hello,
I would like to estimate the parameters of multinomial GLMMs by maximum
likelihood. Or at least, I would like to compute the likelihood of a given
set of parameters.
In the case of multinomial GLMs (i.e. with discrete nominal response
variable), I would use a command like
> library(VGAM)
> summary(vglm(as.factor(output) ~ var1, data = mydataset,
> family="multinomial"))
In the case of binomial GLMMs (i.e. with binary response variable), I would
use a command like
> library(lme4)
> summary(glmer(as.factor(output.bin) ~ var1 + (1 | cluster1), data >
mydataset.bin, family = binomial))
How to compute the likelihood in multinomial GLMMs, then ? (They can be seen
as an extension of multinomial GLMs with random effects, or as an extension
of binomial GLMMs to polytomous data). Or even better: how to estimate the
parameters by maximum likelihood ?
Thanks for your help,
JB
--
View this message in context:
http://r.789695.n4.nabble.com/Multinomial-GLMMs-in-R-tp2955987p2955987.html
Sent from the R help mailing list archive at Nabble.com.