Machine learning or discrete choice models for car ownership demand estimation and prediction?

Paredes, Miguel; Hemberg, Erik; O’Reilly, Una-May; Zegras, Christopher

doi:10.1109/mtits.2017.8005618

Cited by 38 publications

(27 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Understanding mobility choices is key to device low-carbon policies in urban transportation. Estimating the determinants of decisions, ML methods can perform better than standard discrete choice models, for instance in modelling car ownership (Paredes, Hemberg, O'Reilly, & Zegras, 2017). DL (Wang & Zhao, 2018) or semi-supervised Bayesian learning (Yang, Shebalov, & Klabjan, 2018) directly substitute for the traditional logit and probit functions in the model formulation.…”

Section: Engaging With Human Behaviorsmentioning

confidence: 99%

Machine learning for geographically differentiated climate change mitigation in urban areas

Milojevic-Dupont

Creutzig

2021

Sustainable Cities and Society

View full text Add to dashboard Cite

Artificial intelligence and machine learning are transforming scientific disciplines, but their full potential for climate change mitigation remains elusive. Here, we conduct a systematic review of applied machine learning studies that are of relevance for climate change mitigation, focusing specifically on the fields of remote sensing, urban transportation, and buildings. The relevant body of literature spans twenty years and is growing exponentially. We show that the emergence of big data and machine learning methods enables climate solution research to overcome generic recommendations and provide policy solutions at urban, street, building and household scale, adapted to specific contexts, but scalable to global mitigation potentials. We suggest a meta-algorithmic architecture and framework for using machine learning to optimize urban planning for accelerating, improving and transforming urban infrastructure provision. climate change mitigation a distant goal, but also making comparisons between local-scale studies difficult (Lamb, Callaghan, Creutzig, Khosla, & Minx, 2018). The IPCC's AR5 reports knowledge gaps on urban climate action (IPCC, 2014): there is too little understanding of the magnitude of the emissions reductions from altering urban form, and emissions savings from integrated infrastructure and land use planning. New analyses are required both to understand relationships (

show abstract

Section: Engaging With Human Behaviorsmentioning

confidence: 99%

Machine learning for geographically differentiated climate change mitigation in urban areas

Milojevic-Dupont

Creutzig

2021

Sustainable Cities and Society

View full text Add to dashboard Cite

show abstract

“…Individual decision-making has been an important topic in many domains, including marketing [20], economics [33], transportation [5,49], biology [46], and public policy [10]. In recent years as ML models permeated into these domains, researchers started to use various classifiers to analyze how individuals take decisions [39,26]. In the transportation domain, Karlaftis and Vlahogianni (2011) [26] summarized the transportation fields in which DNN models are used, including (1) traffic operations (such as traffic forecasting and traffic pattern analysis); (2) infrastructure management and maintenance (such as pavement crack modeling and intrusion detection); (3) transportation planning (such as in travel mode choice and route choice modeling); (4) environment and transport (such as air pollution prediction); (5) safety and human behavior (such as accident analysis); and (6) air, transit, rail, and freight operations.…”

Section: Literature Reviewmentioning

confidence: 99%

“…For decades, choice analysis has been an important research area across economics, transportation, and marketing [33,5,20]. Whereas discrete choice models were traditionally used to analyze this question, recently researchers have become increasingly interested in applying machine learning (ML) methods such as deep neural network (DNN) to analyze individual choices [26,39,53].…”

Section: Introductionmentioning

confidence: 99%

Deep neural networks for choice analysis: Architecture design with alternative-specific utility functions

Wang

Zhao

2020

Transportation Research Part C: Emerging Technologies

View full text Add to dashboard Cite

“…In studies that compare machine learning algorithms, MNL is typically included as a baseline predictor, with mixed results. There is no clear answer to which algorithm type outperforms others, and the outcome is often dependent on how input variables are structured ( 24 , 25 ). Predicting transportation mode choice using decision tree classification (the method used in this study) has been studied, notably using Random Forest (i.e., tree-based ensemble learning methods) ( 26 , 27 ).…”

Section: Literature Reviewmentioning

confidence: 99%

Attitudes on Autonomous Vehicle Adoption using Interpretable Gradient Boosting Machine

Lee

Mulrow

Haboucha

et al. 2019

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

This article applies machine learning (ML) to develop a choice model on three choice alternatives related to autonomous vehicles (AV): regular vehicle (REG), private AV (PAV), and shared AV (SAV). The learned model is used to examine users’ preferences and behaviors on AV uptake by car commuters. Specifically, this study applies gradient boosting machine (GBM) to stated preference (SP) survey data (i.e., panel data). GBM notably possesses more interpretable features than other ML methods as well as high predictive performance for panel data. The prediction performance of GBM is evaluated by conducting a 5-fold cross-validation and shows around 80% accuracy. To interpret users’ behaviors, variable importance (VI) and partial dependence (PD) were measured. The results of VI indicate that trip cost, purchase cost, and subscription cost are the most influential variables in selecting an alternative. Moreover, the attitudinal variables Pro-AV Sentiment and Environmental Concern are also shown to be significant. The article also examines the sensitivity of choice by using the PD of the log-odds on selected important factors. The results inform both the modeling of transportation technology uptake and the configuration and interpretation of GBM that can be applied for policy analysis.

show abstract

Machine learning or discrete choice models for car ownership demand estimation and prediction?

Cited by 38 publications

References 6 publications

Machine learning for geographically differentiated climate change mitigation in urban areas

Machine learning for geographically differentiated climate change mitigation in urban areas

Deep neural networks for choice analysis: Architecture design with alternative-specific utility functions

Attitudes on Autonomous Vehicle Adoption using Interpretable Gradient Boosting Machine

Contact Info

Product

Resources

About