Abstract Discrete choice models (DCMs) require a priori knowledge of the utility functions, especially how tastes vary across individuals. Utility misspecification may lead to biased estimates, inaccurate interpretations and limited predictability. In this paper, we utilize a neural network to learn taste representation. Our formulation consists of two modules: a neural network (TasteNet) that learns taste parameters (e.g., time coefficient) as flexible functions of individual characteristics; and a multinomial logit (MNL) model with utility functions defined with expert knowledge. Taste parameters learned by the neural network are fed into the choice model and link the two modules. Our approach extends the L-MNL model (Sifringer et al., 2020) by allowing the neural network to learn the interactions between individual characteristics and alternative attributes. Moreover, we formalize and strengthen the interpretability condition — requiring realistic estimates of behavior indicators (e.g., value-of-time, elasticity) at the disaggregated level, which is crucial for a model to be suitable for scenario analysis and policy decisions. Through a unique network architecture and parameter transformation, we incorporate prior knowledge and guide the neural network to output realistic behavior indicators at the disaggregated level. We show that TasteNet-MNL reaches the ground-truth model’s predictability and recovers the nonlinear taste functions on synthetic data. Its estimated value-of-time and choice elasticities at the individual level are close to the ground truth. In contrast, exemplary logit models with misspecified systematic utility lead to biased parameter estimates and lower prediction accuracy. On a publicly available Swissmetro dataset, TasteNet-MNL outperforms benchmarking MNLs and Mixed Logit model’s predictability. It learns a broader spectrum of taste variations within the population and suggests a higher average value-of-time. Our source code is available for research and application.

    Highlights Novel integration of neural networks into discrete choice models (DCM). Neural networks embedded into MNL utility to learn systematic taste heterogeneity. Increased predictability while satisfying more strict interpretability criteria. Incorporation of expert knowledge into network design to achieve interpretability. Experiments on synthetic datasets and the publicly available Swissmetro dataset.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A neural-embedded discrete choice model: Learning taste representation with strengthened interpretability


    Contributors:


    Publication date :

    2022-07-01


    Size :

    21 pages




    Type of media :

    Article (Journal)


    Type of material :

    Electronic Resource


    Language :

    English




    Combining discrete choice models and neural networks through embeddings: Formulation, interpretability and performance

    Arkoudi, Ioanna / Krueger, Rico / Azevedo, Carlos Lima et al. | Elsevier | 2023


    Enhancing discrete choice models with representation learning

    Sifringer, Brian / Lurkin, Virginie / Alahi, Alexandre | Elsevier | 2020


    A DISCRETE CHOICE MODEL WITH TASTE HETEROGENEITY USING SP, RP AND ATTRIBUTE IMPORTANCE RATINGS

    Sasaki, K. / Morikawa, T. / Kawakami, S. et al. | British Library Conference Proceedings | 1999



    Confounding between taste heterogeneity and error structure in discrete choice models

    Hess, S. / Polak, J. / Bierlaire, M. et al. | British Library Conference Proceedings | 2006