Large Language Models (LLMs) are widely used to support various workflows across different disciplines, yet their potential in choice modelling remains relatively unexplored. This work examines the potential of LLMs as assistive agents in the specification and, where technically feasible, estimation of Multinomial Logit models. We implement a systematic experimental framework involving thirteen versions of six leading LLMs (ChatGPT, Claude, DeepSeek, Gemini, Gemma, and Llama) evaluated under five experimental configurations. These configurations vary along three dimensions: modelling goal (suggesting vs. suggesting and estimating MNLs); prompting strategy (Zero-Shot vs. Chain-of-Thoughts); and information availability (full dataset vs. data dictionary only). Each LLM-suggested specification is implemented, estimated, and evaluated based on goodness-of-fit metrics, behavioural plausibility, and model complexity. Findings reveal that proprietary LLMs can generate valid and behaviourally sound utility specifications, particularly when guided by structured prompts. Open-weight models such as Llama and Gemma struggled to produce meaningful specifications. Claude 4 Sonnet consistently produced the best-fitting and most complex models, while GPT models suggested models with robust and stable modelling outcomes. Some LLMs performed better when provided with just data dictionary, suggesting that limiting raw data access may enhance internal reasoning capabilities. Among all LLMs, GPT o3 was uniquely capable of correctly estimating its own specifications by executing self-generated code. Overall, the results demonstrate both the promise and current limitations of LLMs as assistive agents in choice modelling, not only for model specification but also for supporting modelling decision and estimation, and provide practical guidance for integrating these tools into choice modellers' workflows.

Random utility maximisation (RUM) models are one of the cornerstones of discrete choice modelling. However, specifying the utility function of RUM models is not straightforward and has a considerable impact on the resulting interpretable outcomes and welfare measures. In this paper, we propose a new discrete choice model based on artificial neural networks (ANNs) named "Alternative-Specific and Shared weights Neural Network (ASS-NN)", which provides a further balance between flexible utility approximation from the data and consistency with two assumptions: RUM theory and fungibility of money (i.e., "one euro is one euro"). Therefore, the ASS-NN can derive economically-consistent outcomes, such as marginal utilities or willingness to pay, without explicitly specifying the utility functional form. Using a Monte Carlo experiment and empirical data from the Swissmetro dataset, we show that ASS-NN outperforms (in terms of goodness of fit) conventional multinomial logit (MNL) models under different utility specifications. Furthermore, we show how the ASS-NN is used to derive marginal utilities and willingness to pay measures.

Variational Bayes (VB) methods have emerged as a fast and computationally-efficient alternative to Markov chain Monte Carlo (MCMC) methods for Bayesian estimation of mixed multinomial logit (MMNL) models. It has been established that VB is substantially faster than MCMC at practically no compromises in predictive accuracy. In this paper, we address two critical gaps concerning the usage and understanding of VB for MMNL. First, extant VB methods are limited to utility specifications involving only individual-specific taste parameters. Second, the finite-sample properties of VB estimators and the relative performance of VB, MCMC and maximum simulated likelihood estimation (MSLE) are not known. To address the former, this study extends several VB methods for MMNL to admit utility specifications including both fixed and random utility parameters. To address the latter, we conduct an extensive simulation-based evaluation to benchmark the extended VB methods against MCMC and MSLE in terms of estimation times, parameter recovery and predictive accuracy. The results suggest that all VB variants perform as well as MCMC and MSLE at prediction and recovery of all model parameters with the exception of the covariance matrix of the multivariate normal mixing distribution. In particular, VB with nonconjugate variational message passing and the delta-method (VB-NCVMP-Delta) is relatively accurate and up to 15 times faster than MCMC and MSLE. On the whole, VB-NCVMP-Delta is most suitable for applications in which fast predictions are paramount, while MCMC should be preferred in applications in which accurate inferences are most important.

Discrete choice models generally assume that model specification is known a priori. In practice, determiningthe utility specification for a particular application remains a difficult task and model misspecification may lead to biased parameter estimates. In this paper, we propose anew mathematical framework for estimating choice models in which the systematic part of the utility specification is divided into an interpretable part and a learning representation partthat aims at automatically discovering a good utility specification from available data. We show the effectiveness of our framework by augmenting the utility specification of the Multinomial Logit Model (MNL) with a new nonlinear representation arising from a Neural Network (NN). This leads to a new choice model referred to as the Learning Multinomial Logit(L-MNL) model. Our experiments show that our L-MNL model outperformed the traditional MNL models and existing hybrid neural network models both in terms of predictive performance and accuracy in parameter estimation. Keywords: Discrete choice models, Neural networks, Utility specification 1. Introduction Discrete Choice Models (DCM) have emerged as a powerful theoretical framework for analyzing individual travel behavior. The goal of these models is to predict the choice among a given set of discrete alternatives (e.g., choice of walking as transportation mode rather than taking the car or the bus), while understanding the behavioral process that led to the specific choice. For many years, the Multinomial Logit Model (MNL) based on a linear utility specification has provided the foundation for the analysis of discrete choice. Despite its oversimplified assumptions regarding the actual decision-making, this model is still commonly used in practice because it enables a high level of interpretability. Interpretability is critical for researchers and practitioners to get insights into the complex human decision-making process. For instance, linear specifications allow for straightforward derivation of the value-of-time (VOT), i.e., the marginal rate of substitution between time and cost, that constitutes a highly relevant measure in a wide range of public transport policy.