I would like to get the predicted values (with confidence intervals) for a multinomial logistic regression. I know this could be done with predict but in my case I have clustered standard errors in the following way:
multinom <- mlogit(Y ~0| X1+ X2 , data)
cl.mlogit <- function(fm, cluster){
M <- length(unique(cluster))
N <- length(cluster)
K <- length(coefficients(fm))
dfc <- (M/(M-1))
uj <- apply(estfun(fm),2, function(x) tapply(x, cluster, sum));
vcovCL <- dfc*sandwich(fm, meat.=crossprod(uj)/N)
coeftest(fm, vcovCL)
}
cl.mlogit(multinom, data$group)
How I could use these results to get the predicted probabilities (with confidence intervals) for X1=1 and X2=0 for example and compare it with predicted probalities for X1=2 and X2=0. Also, how could I get a confidence interval for that difference? In stata prvalue can do this but don´t know if there is an easy way to do it in R.)
