A literature review of machine learning algorithms for crash injury severity prediction

Santos, Kenny; Dias, João P.; Amado, Conceição

doi:10.1016/j.jsr.2021.12.007

Cited by 78 publications

(23 citation statements)

References 97 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This research analyzed 56 of such research papers published during 2001 to 2021. The review paper reported that RF was the algorithm with the best results, achieving the best performance in 70% of the times that it was used, and in 29% of all the inspected studies [ 73 ]. Hence, the existing methodological reasons and empirical evidence in the literature verify the obtained results in this research.…”

Section: Discussionmentioning

confidence: 99%

Classification of truck-involved crash severity: Dealing with missing, imbalanced, and high dimensional safety data

2023

View full text Add to dashboard Cite

While the cost of road traffic fatalities in the U.S. surpasses $240 billion a year, the availability of high-resolution datasets allows meticulous investigation of the contributing factors to crash severity. In this paper, the dataset for Trucks Involved in Fatal Accidents in 2010 (TIFA 2010) is utilized to classify the truck-involved crash severity where there exist different issues including missing values, imbalanced classes, and high dimensionality. First, a decision tree-based algorithm, the Synthetic Minority Oversampling Technique (SMOTE), and the Random Forest (RF) feature importance approach are employed for missing value imputation, minority class oversampling, and dimensionality reduction, respectively. Afterward, a variety of classification algorithms, including RF, K-Nearest Neighbors (KNN), Multi-Layer Perceptron (MLP), Gradient-Boosted Decision Trees (GBDT), and Support Vector Machine (SVM) are developed to reveal the influence of the introduced data preprocessing framework on the output quality of ML classifiers. The results show that the GBDT model outperforms all the other competing algorithms for the non-preprocessed crash data based on the G-mean performance measure, but the RF makes the most accurate prediction for the treated dataset. This finding indicates that after the feature selection is conducted to alleviate the computational cost of the machine learning algorithms, bagging (bootstrap aggregating) of decision trees in RF leads to a better model rather than boosting them via GBDT. Besides, the adopted feature importance approach decreases the overall accuracy by only up to 5% in most of the estimated models. Moreover, the worst class recall value of the RF algorithm without prior oversampling is only 34.4% compared to the corresponding value of 90.3% in the up-sampled model which validates the proposed multi-step preprocessing scheme. This study also identifies the temporal and spatial (roadway) attributes, as well as crash characteristics, and Emergency Medical Service (EMS) as the most critical factors in truck crash severity.

show abstract

Section: Discussionmentioning

confidence: 99%

Classification of truck-involved crash severity: Dealing with missing, imbalanced, and high dimensional safety data

2023

View full text Add to dashboard Cite

show abstract

“…In recent years, it is universally employed in scientific research related to major trauma. For instance, Versluijs et al ( 62 ) reviewed the association between trauma severity and post-injury symptoms of depression; Santos et al ( 63 ) predicted the severity of crash injury by investigating machine learning algorithms.…”

Section: Discussionmentioning

confidence: 99%

Dynamic structures and emerging trends in the management of major trauma: A bibliometric analysis of publications between 2012 and 2021

Wang²,

Guo³

et al. 2022

Front. Public Health

View full text Add to dashboard Cite

ObjectiveMajor trauma is currently a global public health issue with a massive impact on health at both the individual and population levels. However, there are limited bibliometric analyses on the management of major trauma. Thus, in this study we aimed to identify global research trends, dynamic structures, and scientific frontiers in the management of major trauma between 2012 and 2021.MethodsWe searched the Web of Science Core Collection to access articles and reviews concerning the management of major traumas and conducted a bibliometric analysis using CiteSpace.ResultsOverall, 2,585 studies were screened and published by 403 institutions from 110 countries/regions. The most productive country and institution in this field of research were the USA and Monash University, respectively. Rolf Lefering was the most prolific researcher and Holcomb JB had the most co-citations. Injury published the highest number of articles, and the Journal of Trauma was the most co-cited journal. A dual-map overlay of the literature showed that the articles of most publications were confined to the areas of medicine/medical/clinical and neurology/sports/ophthalmology. Document clustering indicated severe traumatic brain injury, traumatic coagulopathy, and resuscitative endovascular balloon occlusion as the recent hot topics. The most recent burst keywords were “trauma management,” “neurocritical care,” “injury severity,” and “emergency medical services.”ConclusionThe dynamic structures and emerging trends in the management of major trauma were extensively analyzed using CiteSpace, a visualization software. Based on the analysis, the following research hotspots emerged: management of severe traumatic brain injury and massive hemorrhage, neurocritical care, injury severity, and emergency medical service. Our findings provide pertinent information for future research and contribute toward policy making in this field.

show abstract

“…These researchers proved that the fuzzy algorithms proved high prediction accuracy for RTAs occurrence in the studied locations. Detailed review on the applications of machine learning in RTAs models can be found in Tang et al [23] and Santos et al [36]. This paper is concerned with modeling the relationship between accident number and the influencing factors using fuzzy logic algorithm to develop relationships between multiple inter-related factors affecting the RTAs.…”

Section: Literature Reviewmentioning

confidence: 99%

Analysis And Modeling of Rural Roads Traffic Safety Data

Gaber

Diab²,

Othman³

et al. 2023

Sohag Engineering Journal

View full text Add to dashboard Cite

Traffic accidents threaten human's safety and properties as well as restrict the management of transportation systems. Therefore, traffic accidents prediction models could help stakeholders to understand accident causes toward controlling their effects. This paper aims at developing a prediction model for a rural road in Egypt using actual accident data with survey data for pavement conditions, traffic flow presented as average hourly traffic per lane, speed, minor access, traffic signs conditions and road width. The relationship between accident number and the influencing factors is modeled using fuzzy logic algorithm which deals with nonlinear relationships between parameters. The results show that the developed model produced accurate and stable traffic accident predictions (R 2 =0.79). The results indicated that average accident rate decreases as road width increases where, a reduction in accident rate of about 5.26 % is reached due to road width widening by 3.5 m. The results indicate that road accidents on the studied road can be reduced by 7.38% if the speed is limited to 75 km per hour. About 17.90% reduction is achieved by improving road surface condition. Also results indicate a positive relationship between traffic volume and accidents rate.

show abstract

A literature review of machine learning algorithms for crash injury severity prediction

Cited by 78 publications

References 97 publications

Classification of truck-involved crash severity: Dealing with missing, imbalanced, and high dimensional safety data

Classification of truck-involved crash severity: Dealing with missing, imbalanced, and high dimensional safety data

Dynamic structures and emerging trends in the management of major trauma: A bibliometric analysis of publications between 2012 and 2021

Analysis And Modeling of Rural Roads Traffic Safety Data

Contact Info

Product

Resources

About