A Customized Deep Neural Network Approach to Investigate Travel Mode Choice with Interpretable Utility Information

Abstract

Discrete choice modeling of travel modes is an essential part of traffic planning and management. Thus far, this field has been dominated by multinomial logit (MNL) models with a linear utility specification. However, deep neural networks (DNNs), owing to their powerful capacity of nonlinear fitting, are now rapidly replacing these models. This is because, by using DNNs, mode choice can be assimilated with the classification problems within the machine learning community. This article proposes a newly designed DNN framework for traffic mode choice in the style of two hidden layers. First, a local-connected layer automatically extracts an effective utility specification from the available data, and then, a fully connected layer augments the feature representation. Validated by a practical city-wide multimodal traffic dataset in Beijing, our model significantly outperforms the random utility models and simple fully connected neural network in terms of the prediction accuracy. Besides the comparison of the predictive power, we also present the interpretability of the proposed model.

Analysis of Travel Mode Choice in Seoul Using an Interpretable Machine Learning Approach

“…Comparing the interpretation results of ML models with an advanced parametric model, such as a mixed logit model, would also be valuable to validate the model further. Deep learning models [11,42] are reasonable alternatives for the XGB and RF and the proposed modelagnostic interpretation methods can still available for those models. Local interpretation methods such as local interpretable model-agnostic explanations (LIME) and Shapley additive explanations (SHAP) can contribute to better representation of the heterogeneity of individuals and groups [43,44], which has also been a critical subject of behavior analysis.…”

“…It can represent complex relationships between mode choices and input variables in a data-driven manner rather than making strict assumptions about the data [7]. Many previous studies have reported the use of an ML approach to model travel mode choice [1,[6][7][8][9][10][11]. ese authors have generally reported improvements in the prediction performance of ML approaches compared to MNLbased models.…”

An Effective Approach to Promote Air Traveler Repurchasing Using the Random Forest Algorithm: Predictive Model Design and Utility Evaluation

“…ere is much literature on predicting air traveler repeated purchase behaviors [8,10,14,[18][19][20], but literature based on using a machine learning model to make these predictions is still limited [8,12,15]. is is especially true for actual traveler sales date, because surveys or experimental data have often been used to study traveler repeat purchases [10,18,22,26].…”

“…is is especially true for actual traveler sales date, because surveys or experimental data have often been used to study traveler repeat purchases [10,18,22,26]. Besides, some studies have optimized other machine learning algorithms, such as artificial bee colony algorithm and deep neural network approach, to predict customer repurchase intention [12,15]. However, these studies only use consumer data in specific environments to test their proposed optimization algorithms and do not study whether these algorithms can be well applied to the business environment.…”

“…Machine learning is an important data-mining technique. Machine learning and data-mining techniques for predictive purposes on an air traveler basis are often applied in air traveler relationship management and expert systems domains [14,15], and predicting air traveler repurchases is a popular objective in these fields. In order to identify travelers with high purchase intention, in this study, the existing machine learning model has been proposed to predict the purchase intention of travelers in airlines.…”