Machine learning for traffic crash prediction: a systematic review of performance and interpretability
SAP Transport, Mobility & Society · 23.09.2026 · 5 · 327
DOI: 10.62486/tms2026327 ↗Kurz gefasst
Systematische Übersicht in der Zeitschrift SAP Transport, Mobility & Society, online erschienen am 23.09.2026. Verfasst von Carlos Ignacio Ruiz Maldonado und 1 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Introduction
Road traffic crashes cause more than 1.1 million deaths each year, and machine learning is increasingly used to predict where and how severe they will be.
Methods
A PRISMA 2020 systematic review was conducted of empirical studies published between 2013 and 2026 that applied machine learning to traffic crash data.
Results
From 375 records, 298 were screened and 19 studies were included. Most studies targeted crash severity (14) rather than occurrence or frequency (5), and tree ensembles dominated. SHAP was used in 13 studies, and external validation in only 1. Overall risk of bias was low in 11 studies, moderate in 5, and high in 3.
Conclusions
Machine learning predicts crash severity with high but variable accuracy, interpretability is now common, and external validation remains rare and should become standard before deployment.
Zitieren
APA
Maldonado, C. I. R., & Roman-Acosta, D. (2026). Machine learning for traffic crash prediction: a systematic review of performance and interpretability. SAP Transport, Mobility & Society, 5, 327. https://doi.org/10.62486/tms2026327
Vancouver
Maldonado CIR, Roman-Acosta D. Machine learning for traffic crash prediction: a systematic review of performance and interpretability. SAP Transport, Mobility & Society. 2026;5:327. doi:10.62486/tms2026327
BibTeX
@article{Maldonado2026,
author = {Maldonado, Carlos Ignacio Ruiz and Roman-Acosta, Daniel},
title = {Machine learning for traffic crash prediction: a systematic review of performance and interpretability},
journal = {SAP Transport, Mobility & Society},
year = {2026},
volume = {5},
pages = {327},
doi = {10.62486/tms2026327}
}
EUVIMED