Citation

Leal JE, Parada V. Transp. Res. Interdiscip. Persp. 2023; 22: e100947.

Copyright

(Copyright © 2023, Elsevier Publishing)

DOI

10.1016/j.trip.2023.100947

PMID

unavailable

Abstract

Demand estimation and forecasting is an essential step in urban passenger transport planning. Relating the factors that influence the modal choice behavior of individuals facilitates demand estimation. In this study, we develop machine learning models that consider individuals' demographic, socioeconomic, and travel characteristics to justify their mode choice. Two datasets are used to train and validate the models. We use logistic regression and multilayer perceptron models to classify public or private transportation trips. It was observed that a multilayer perceptron model with a low number of parameters could predict modal selection with an accuracy exceeding 90%. We derive an algebraic equation from this result to perform modal selection prediction. Our results show that the models can effectively predict the mode of transportation of individuals based on their demographic and travel characteristics.

Language: en

Keywords

Machine learning; Modal choice; Transportation demand