I am trying to estimate a multinomial logit model of customer store choice.
There are different choice sets for each customer. The dataset is large, so mlogit take too long to converge. mnlogit is faster with large data sets, but I can't figure out how to specify that the choice sets vary. Is there an easy fix that I am missing? The dataset is a mlogit.data object, and I have set the chid.var setting to a choice index variable.
I am not stuck on using mnlogit, but it seems like my best option given that I have a large dataset. I can't find any mention of differing choice sets in the mnlogit documentation.
Edit:
The warning I get with mnlogit:
In responseMat - Pch : longer object length is not a multiple of shorter object length
