artículo
Inclusion of latent variables in Mixed Logit models: Modelling and forecasting
Fecha
2010Registro en:
10.1016/j.tra.2010.07.007
0965-8564
s2.0-77956318583
WOS:000282485400007
Autor
Yáñez Castillo, María Francisca
Raveau Feliú, Sebastián
Ortúzar Salas, Juan de Dios
Institución
Resumen
Travel demand models typically use mainly objective modal attributes as explanatory variables. Nevertheless, it has been well known for many years that attitudes and perceptions also influence users' behaviour. The use of hybrid discrete choice models constitutes a good alternative to incorporate the effect of subjective factors. We estimated hybrid models in a short-survey panel context for data among many alternatives. The paper analyses the results of applying these models to a real urban case study, and also proposes an approach to forecasting using these models. Our results show that hybrid models are clearly superior to even highly flexible traditional models that ignore the effect of subjective attitudes and perceptions.