Enhancing discrete choice models with representation learning. (October 2020)

• Record Type: Journal Article
• Authors: Sifringer, Brian; Lurkin, Virginie; Alahi, Alexandre
• Is Part Of: Transportation research. Volume 140(2020)
• Page Start: 236
• Page End: 261
• Publication Date: 2020-10
• Subjects: Discrete choice models -- Neural networks -- Utility specification -- Machine learning -- Deep learning; Transportation -- Research -- Periodicals; Transportation -- Mathematical models -- Periodicals
• DOI: 10.1016/j.trb.2020.08.006
• Languages: English
• ISSNs: 0191-2615

Abstract

Highlights: New data-driven framework for enhancing choice models. Conservation of standard DCM interpretability while increasing predictive power. Systematic utility divided into a knowledge-driven and a data-driven part. Demonstration of framework's effectiveness on the MNL and NL models. New choice models referred to as the Learning Multinomial Logit (L-MNL) and Learning Nested Logit (L-NL) models. Experiments on publicly available datasets based on revealed or stated preferences. Abstract: In discrete choice modeling (DCM), model misspecifications may lead to limited predictability and biased parameter estimates. In this paper, we propose a new approach for estimating choice models in which we divide the systematic part of the utility specification into (i) a knowledge-driven part, and (ii) a data-driven one, which learns a new representation from available explanatory variables. Our formulation increases the predictive power of standard DCM without sacrificing their interpretability. We show the effectiveness of our formulation by augmenting the utility specification of the Multinomial Logit (MNL) and the Nested Logit (NL) models with a new non-linear representation arising from a Neural Network (NN), leading to new choice models referred to as the Learning Multinomial Logit (L-MNL) and Learning Nested Logit (L-NL) models. Using multiple publicly available datasets based on revealed and stated preferences, we show that our models outperform the traditional ones, both in terms of predictive performance and accuracy in parameter estimation. All source code of the models are shared to promote open science.