-
Explain why mixed logit is such a popular model, according to Train. What are its strong points vis-à-vis the IIA problem?
-
What are the computational drawbacks to mixed logit?
-
What are two names for a mixed logit model that has a discrete mixing distribution?
-
Explain why the EM algorithm is a potentially powerful tool for estimating discrete choice models.
-
How does the EM algorithm work? Why does it work?