Van Cranenburgh and Kouwenhoven (under review) propose a new nonparametric method to uncover the Value-of-Travel-Time (VTT) distribution using data from a binary within-mode experiment (in which decision makers have to choose between a slow and cheap alternative and a fast and expensive alternative). In the core of this method – which we henceforth refer to as CK method – is an Artificial Neural Network (ANN). The ANN captures associations between the tuple of explanatory variables (the series of T choices, the received Boundary Value-of-Travel-Times (BVTTs), and a set of socio-demographic variables), and the dependent variable: the probability of choosing the fast and expensive alternative in the ‘next’ T+1th choice task. The CK method is appealing as it uncovers the VTT distribution (and its moments) without making strong assumptions on the underlying behaviour. Moreover, unlike other nonparametric approaches the method incorporates covariates, accounts for panel effects and yields a distribution right of the maximum received BVTT.
Original languageEnglish
PublisherDelft University of Technology
Publication statusIn preparation - 2019

ID: 53863102