15h30 - 15h55
Implementation of a Maximum Simulated Likelihood Estimator for a Mixed Logit Model with Endogenous Latent Explanatory Variables
In this paper we expand on our previous research on hybrid choice models that
incorporate latent explanatory variables by considering a flexible choice kernel
that allows for random heterogeneity. In particular, we analyze the case of a
mixed logit model for the discrete-choice kernel of the hybrid choice model. We
implement a maximum simulated likelihood estimator and, using a real dataset
on travel mode choice, we test the performance of this estimator for a crossnested
structure as well as interactions between the latent variables and the
exogenous attributes in the utility function. Our study suggests that the estimator
we implemented is valid and appropriate for empirical applications.
15h55 - 16h20
Latent Segmentation Based Discrete Choice Models
In this paper, we aim to identify the different factors that influence injury severity of highway vehicle occupants, in particular drivers, involved in a vehicle-train collision at highway-railway grade crossings. The commonly used approach to modeling vehicle occupant injury severity is the traditional ordered response model that assumes the effect of various exogenous factors on injury severity to be constant across all accidents. The current research effort attempts to address this issue by applying an innovative latent segmentation based ordered logit model to evaluate the effects of various factors on the injury severity of vehicle drivers. In this model, the highway-railway crossings are assigned probabilistically to different segments based on their attributes with a separate injury severity component for each segment. The validity and strength of the formulated collision consequence model is tested using the US Federal Railroad Administration database which includes inventory data of all the railroad crossings in the US and collision data at these highway railway crossings from 1997 to 2006. The model estimation results clearly highlight the existence of risk segmentation within the affected grade crossing population by the presence of active warning devices, presence of permanent structure near the crossing and roadway type. The key factors influencing injury severity include driver age, time of the accident, presence of snow and/or rain, vehicle role in the crash and motorist action prior to the crash.
16h20 - 16h45
Scale Heterogeneity in Choice Models: Capturing Absolute Heterogeneity in Discrete Choice Models
Heterogeneity in choice model is a serious issues and long been under the focus of many researchers. Heterogeneity in choice model is classified into two categories: relative heterogeneity and absolute heterogeneity. Relative heterogeneity is captured by attributes parameters, choice model formulation, error structure etc. On the other hands, the absolute heterogeneity is captured by scale parameterization of econometric choice models. In typical logit model formulation scale heterogeneity is completely overlooked because of identification issue. In advanced mixed logit model structures, main focuses are on capturing the relative heterogeneity by attribute correlations. Compared to relative heterogeneity, the scale heterogeneity has received less focus in empirical and theoretical exercises. Recently there has been an increasing interest in capturing scale heterogeneity in the choice models as the parameter heterogeneities are not always enough in explaining choice variations. However, capturing scale heterogeneity is not an easy task as it relates to the model identification problems. There are conflicting findings and arguments on our capacity to separate scale heterogeneity from parameter heterogeneity. However, it is the general consensus that capturing scale heterogeneity is a very important issue. In this presentation, I would present my recent works on capturing scale heterogeneity in econometric choice models. My presentation will mostly focus on the importance and feasibility of explaining scale heterogeneity systematically so that the application of the models remain straight forward for practical policy analysis.