Exploiting Machine Learning Algorithms for Predicting Crash Injury Severity in Yemen: Hospital Case Study

Abstract

This study focused on exploiting machine learning algorithms for classifying and predicting injury severity of vehicle crashes in Yemen. The primary objective is to assess the contribution of the leading causes of injury severity. The selected machine learning algorithms compared with traditional statistical methods. The filtrated second data collected within two months (August-October 2015) from the two main hospitals included 156 injured patients of vehicle crashes reported from 128 locations. The data classified into three categories of injury severity: Severe, Serious, and Minor. It balanced using a synthetic minority oversampling technique (SMOTE). Multinomial logit model (MNL) compared with five machine learning classifiers: Naïve Bayes (NB), J48 Decision Tree, Random Forest (RF), Support Vector Machine (SVM), and Multilayer Perceptron (MLP). The results showed that most of machine learning-based algorithms performed well in predicting and classifying the severity of the traffic injury. Out of five classifiers, RF is the best classifier with 94.84% of accuracy. The characteristics of road type, total injured person, crash type, road user, transport way to the emergency department (ED), and accident action were the most critical factors in the severity of the traffic injury. Enhancing strategies for using roadway facilities may improve the safety of road users and regulations.

Modeling Crashes Severity Using Ensemble Techniques

“…Results indicated that the various used algorithms have different accuracy rate at different roadway type. Most importantly, their study showed that holidays affect the crash severity (Almannaa et al, 2023).Several works were conducted to model traffic crashes globally ( (Almamlook et al, 2019;Almannaa et al, 2023;Al-Mistarehi et al, 2022;Al-Moqri et al, 2020;Alrumaidhi & Rakha, 2022;Anderson & Hernandez, 2017;Azhar et al, 2022) Most of these studies used machine learning algorithms including K-nearest neighborhoods (KNN), support vector machine, adaptive boosting tree or other machine learning algorithms (Ahmed et al, 2021;Almamlook et al, 2019;Almannaa et al, 2023a;Al-Mistarehi et al, 2022;Al-Moqri et al, 2020). Researchers indicated that providing driver sociodemographic attributes may improve the model prediction ( (Almannaa et al, 2023;Jamal et al, 2021;R.…”

A data-mining study on the prediction of head injury in traffic accidents among vulnerable road users with varying body sizes and head anatomical characteristics

“…Classical machine learning (ML) algorithms (e.g., logistic regression, k-nearest-neighbors, support vector machines, decision tree, random forests, etc.) have been used to develop a prediction model of human injury degree ( Al-Moqri et al, 2020 ; Mansoor et al, 2020 ; Liu et al, 2022 ). At the same time, in order to further explore the interaction relationship between objective factors in traffic accidents and human injury response to improve the accuracy of prediction models, deep learning (DL) method with higher computational complexity (e.g., deep neural network, convolutional neural network, long short-term memory, recurrent neural network, etc.)…”

Predicting and Analyzing Road Traffic Injury Severity Using Boosting-Based Ensemble Learning Models with SHAPley Additive exPlanations

“…Similarly, combining modern data sources frequently produces complex data sets with a large number of dimensions that are difficult to model using traditional statistical techniques. Machine learning methods, on the other hand, are highly adaptable, require few or no prior assumptions about the data, and can handle missing values, noise, and outliers [ 27 ]. In the field of traffic safety research, machine learning techniques have received a lot of attention.…”