Bayesian Deep Learning for Discrete Choice

Abstract

Discrete choice models are used to analyze individual decision-making in contexts such as consumer behavior, transportation choices, and political elections. Whereas traditional discrete choice models offer high economic interpretability, these models often underperform in predictive tasks compared to deep learning. Deep learning models remain largely underutilized in discrete choice due to concerns about their lack of interpretability, unstable parameter estimates, and the absence of established methods for uncertainty quantification. In this work, we introduce a deep learning model architecture that exploits approximate Bayesian inference methods via Stochastic Gradient Langevin Dynamics. Our proposed model collapses to behaviorally informed hypotheses when data is limited, mitigating overfitting and instability in underspecified settings while retaining the flexibility to capture complex nonlinear relationships when sufficient data is available. We demonstrate the value of our approach in a Monte Carlo simulation study, evaluating both predictive metrics –such as out-of-sample balanced accuracy– and inferential metrics –such as empirical coverage for marginal rates of substitution interval estimates. Additionally, we present results from two empirical case studies: one using revealed mode choice data in NYC, and the other based on the widely used Swiss train choice stated preference data.