Travel Mode Choice Modeling: Predictive Efficacy between Machine Learning Models and Discrete Choice Model



Nur Fahriza Mohd Ali1, *, Ahmad Farhan Mohd Sadullah1, Anwar PP Abdul Majeed2, Mohd Azraai Mohd Razman2, Muhammad Aizzat Zakaria2, Ahmad Fakhri Ab. Nasir3
1 School of Civil Engineering, Engineering Campus, Universiti Sains Malaysia, 14300, Nibong Tebal, Pulau Pinang, Malaysia
2 Innovative Manufacturing, Mechatronics and Sports Laboratory, Faculty of Manufacturing and Mechatronics Engineering Technology, Universiti Malaysia Pahang, 26600, Pekan, Pahang, Malaysia
3 Faculty of Computing, College of Computing and Applied Sciences, Universiti Malaysia Pahang, 26600, Pekan, Pahang, Malaysia

Abstract

Background:

A complex travel behaviour among users is intertwined with many factors. Traditionally, the exploration in travel mode choice modeling has been dominated by the Discrete Choice model, nonetheless, owing to the advancement in computational techniques, machine learning has gained traction in understanding travel behavior.

Aim:

This study aims at predicting users’ travel model choice by means of machine learning models against a conventional Discrete Choice Model, i.e., Binary Logistic Regression.

Objective:

To investigate the comparison between machine learning models namely Neural Network, Random Forest, Decision Tree, and Support Vector Machine against the Discrete Choice Model (Binary Logistic Regression) in the prediction of travel mode choice amongst Kuantan City.

Methodology:

The dataset were collected in Kuantan City, Malaysia, through the Revealed/Stated Preferences (RP/SP) Survey. The data collected was split into a ratio of 80:20 for training and testing before evaluating them between the aforesaid models. The hyperparameters of the models were set to default. The performance of the models is evaluated based on the classification accuracy.

Results:

It was shown in the present study that the Neural Network Model is able to attain a higher prediction accuracy as compared to Binary Logistic Regression (Discrete Choice Model) in classifying mode choice of Kuantan users either to choose public transport or private vehicles as daily transportation. Feature importance technique is crucial to identify the significant features in modelling travel mode choice. It is demonstrated that the Neural Network model can yield exceptional classification of mode choice up to 73.4% and 72.4% of training and testing data, respectively, by considering the features identified via the feature importance technique, suggesting the viability of the proposed technique in supporting granting an informed decision.

Conclusions;

The findings highlight the strengths and limitations of the Machine Learning technique as well as the Discrete Choice model in modeling travel mode choice. It was shown that Machine Learning models have the capability in providing better prediction that could assist the urban transportation planning among policymakers. Meanwhile, it could be also demonstrated that the Discrete Choice Model (Binary Logistic Regression) is helpful to get a better understanding in expressing the inference relationship between variables for improvising the future transportation system.

Keywords: Travel mode choice, Revealed/Stated preference (RPSP) survey, Public transport, Private vehicles, Machine learning technique, Discrete choice model.


Abstract Information


Identifiers and Pagination:

Year: 2021
Volume: 14
Publisher Item Identifier: EA-TOTJ-2021-13

Article History:

Electronic publication date: 27/8/2021
Collection year: 2021

© 2021 Ali et al.

open-access license: This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International Public License (CC-BY 4.0), a copy of which is available at: https://creativecommons.org/licenses/by/4.0/legalcode. This license permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


* Address correspondence to this author at School of Civil Engineering, Engineering Campus, Universiti Sains Malaysia, 14300, Nibong Tebal, Pulau Pinang, Malaysia; E-mail: fahriza90@gmail.com