A Dirichlet process mixture model of discrete choice: Comparisons and a case study on preferences for shared automated vehicles

This paper i) compares parametric and semi-parametric representations of unobserved heterogeneity in hierarchical Bayesian logit models and ii) applies these methods to infer distributions of willingness to pay for features of shared automated vehicle (SAV) services. Specifically, we compare the multivariate normal, the finite mixture of normals and the Dirichlet process mixture of normals (DP-MON) mixing distributions. The latter promises to be particularly flexible regarding the shapes it can assume, and unlike other semi-parametric approaches does not require that its complexity is fixed before estimation. We evaluate the different mixing distributions, using simulated data and real data from a stated choice study on preferences for SAVs in New York City. In the considered data settings, the DP-MON mixing distribution provides an excellent data fit and performs at least as well as the other methods at out-of-sample prediction. The DP-MON mixing distribution also offers substantive behavioural insights into the adoption of SAVs. We find that preferences for in-vehicle travel time by SAV with ride-pooling are strongly polarised. Whereas one-third of the sample is willing to pay between 10 and 80 USD/h to avoid pooling a vehicle with strangers, the remainder of the sample is either indifferent to ride-pooling or even desires it. We also estimate that vehicle automation and powertrain electrification are relatively unimportant to travellers. Consequently, travellers may primarily derive indirect, rather than immediate benefits from these new technologies through increases in operational efficiency and lower operating costs.