Skip to main content
6 events
when toggle format what by license comment
Jan 14, 2015 at 18:00 comment added gregmacfarlane You're right that the IIA isn't always justified, and you're also right that with modern estimators probit models can be estimated reasonably quickly. But GEV models resolve the IIA problem and might better represent the choice structure in certain situations. I'm also not sure that probit is "more used today;" in my field (transportation modeling), probit models remain a novelty.
Jan 10, 2015 at 21:13 comment added kjetil b halvorsen But, in the choice situation, probit is more flexible, so moore used today! multinomial logit implies the assumption of irrelevance of irrelevant alternatives, which not always is empirically justified.
Sep 28, 2012 at 2:18 comment added gregmacfarlane How would you incorporate spatial elements into a DC model? I'm very interested.
Sep 27, 2012 at 23:37 comment added fmark This is why multinomial logit functions are classically used to estimate spatial discrete choice problems, even though the actual phenomenon is better modelled by a probit.
Jun 23, 2012 at 1:40 history edited gregmacfarlane CC BY-SA 3.0
clarified points, and added some math.
Jun 22, 2012 at 14:29 history answered gregmacfarlane CC BY-SA 3.0