The multinomial logit model is a kind of model which has both alternative variant and invariant independent variables. In such a case, there is a requirement to change into a long format using data frame syntax due to variables nature.
But, I want to use R commander(Rcmdr) to calibrate multinomial logit model. how can I consider these alternative variant and invariant variables?
Alternative specific(alternative variant) variables are: act+wtt+invtt+lact+TC
Generic(alternative invariant) Variables : factor(GEN)+AGE+factor(OCC)+factor(INC)+factor(MAR)+factor(OWN)+factor(Routesused)+DIST...
model 1 <-mlogit(mode~ act+wtt+invtt+lact+TC | factor(GEN)+AGE+factor(OCC)+factor(INC)+factor(MAR)+factor(OWN)+factor(Routesused)+DIST , data= mydata,reflevel="Taxi")
From the above model,we can use "|" sign to separate alternative variant and invariant variables in normal calibration using codes.
My question is how can I employ the variables in r-commander blog which is in the figure below(please , see the attached photo)