On Modeling and Estimation of Travel Behavior Utilizing Varying Contextual Information
Abstract
This paper attempts to outline the essence of a methodological approach which allows the researcher to utilize relevant contextual information in modeling and estimation of travel behavior in a manner that takes account of the varying context. For this purpose, we utilize a varying coefficients model to explain travel behavior, using cross-country data from time-use studies. This study goes beyond previous studies in two directions. First, it examines travel behavior as occurring in a broader context by considering a number of contextual elements which take on meaning within the activity systems framework. Second, as far as our model and test procedure are concerned, our analysis adopts a model which can more correctly relate the response variable to its determinants and also permits the impact of those determinants to be country/group specific. Our empirical results appear to indicate that travel behavior is the product of a variety of forces captured in several contextual dimensions.