Tabular Foundation Models for Discrete Choice Estimation

arXiv cs.AI · 18 hours ago

Researchers developed a method to apply tabular foundation models to discrete choice estimation by reformulating the problem to capture choice-set dependence and consumer preference heterogeneity. The best reformulation outperformed hierarchical Bayesian estimation by 8% in holdout log-likelihood and 3.6% in hit rate while running 16 times faster on yogurt scanner panel data. This approach enables more efficient demand estimation at scale, particularly benefiting scenarios with 10-40 purchase occasions per consumer where traditional parametric Bayesian methods distort estimates.