Abstract:Logit models are usually applied when studying individual travel behavior, i.e., to predict travel mode choice and to gain behavioral insights on traveler preferences. Recently, some studies have applied machine learning to model travel mode choice and reported higher out-of-sample predictive accuracy than traditional logit models (e.g., multinomial logit). However, little research focuses on comparing the interpretability of machine learning with logit models. In other words, how to draw behavioral insights from the high-performance "black-box" machine-learning models remains largely unsolved in the field of travel behavior modeling. This paper aims at providing a comprehensive comparison between the two approaches by examining the key similarities and differences in model development, evaluation, and behavioral interpretation between logit and machine-learning models for travel mode choice modeling. To complement the theoretical discussions, the paper also empirically evaluates the two approaches on the stated-preference survey data for a new type of transit system integrating high-frequency fixed-route services and ridesourcing. The results show that machine learning can produce significantly higher predictive accuracy than logit models. Moreover, machine learning and logit models largely agree on many aspects of behavioral interpretations. In addition, machine learning can automatically capture the nonlinear relationship between the input features and choice outcomes. The paper concludes that there is great potential in merging ideas from machine learning and conventional statistical methods to develop refined models for travel behavior research and suggests some new research directions.

Comparative Study of Logit and Weibit Model in Travel Mode Choice

The Analysis of Dynamic Travel Mode Choice: A Heterogeneous Hidden Markov Approach

Travel Mode Choice Model Accounting for Individual Preference Heterogeneity and Correlation among Choice Alternatives

Analysis and Application for Urban Inhabitant Travel Demand Models

Modeling traveler choice behavior using the concepts of relative utility and relative interest

Modeling Intercity Travel Mode Choice with Data Balance Changes: A Comparative Analysis of Bayesian Logit Model and Artificial Neural Networks

Modelling Travellers’ Route Choice Behaviours with the Concept of Equivalent Impedance

Development of a Multinomial Logit Model for Trave Mode Choice of Residents

On the Development of a Semi-Nonparametric Generalized Multinomial Logit Model for Travel-Related Choices.

Comparison of Mode Split Model Based on Multinomial Logit and Nested Logit

A Study on Intercity Transport System Between Guangzhou and Wuhan Based on Two Logit Models

Modeling Stated Preference for Mobility-on-Demand Transit: A Comparison of Machine Learning and Logit Models

Travel Space and Mode Combined Choice Model for Urban Travelers

Effects of Cost Adjustment on Travel Mode Choice: Analysis and Comparison of Different Logit Models

Study on the Individual Travel Mode Choice Based on the Discrete Choice Theory

A Nested Logit Model for Passenger Travel Behavior of Dedicated Passenger Railway

Investigating Mode Preference of Intercity Business Travelers in the Urban Agglomeration of China: Application of Fractional Multinomial Logit Model

Cross-nested logit model for the joint choice of residential location, travel mode, and departure time

Study of Decision Tree Structure of Traffic Mode Choice Based on the Goodness of Fit Measure

Typical Combined Travel Mode Choice Utility Model in Multimodal Transportation Network

Modeling Stochastic Route Choice Behaviors with Equivalent Impedance