The paired combinatorial logit model: properties, estimation and application

摘要: The Independence of Irrelevant Alternatives (IIA) property the multinomial logit (MNL) model imposes restriction zero covariance between utilities pairs alternatives. This is inappropriate for many choice situations; those in which some or sets alternatives share same unobserved attributes. nested (NL) relaxes MNL but equal among all a common nest and otherwise. paired combinatorial (PCL) these restrictions further by allowing different covariances each pair relaxation enables estimation differential competitive relationships closed form PCL retains computational advantages other models while more flexible error correlation structure, compared to NL models, better representation situations. paper describes derivation, properties model. empirical results demonstrate that statistically superior may lead importantly travel forecasts policy decisions.