Browsing by Author "الضاري، ابراهيم"

Now showing 1 - 1 of 1

Restricted
Severity Prediction Model for Highway Crashes in Qassim, Saudi Arabia Using Machine Learning Algorithms
(Qassim University, 2023-06-01) الضاري، ابراهيم; Aldhari, Ibrahim; المشيقح, مشعل; Almoshaogeh, Meshal
Among the G20 countries, Saudi Arabia is focusing very much on traffic safety. Driver distraction is the primary cause of increased high-severity traffic accidents. Several models including lots of factors have been presented in investigations of severity of traffic accidents world-wide. Most of these studies are data specific and region specific. there are so many methods of analysis for this issue that the topic will remain hot issue for coming few years. In this study, severity prediction models were developed and implemented for Qassim Province in Saudi Arabia. Traffic accident data for the assessment period from January 2017 to December 2019 were obtained from the Ministry of Transport and Logistic Services. Three classifiers, two of which are ensemble machine learning methods including random forest and XGBoost, and third being logistic regression, modeled the crash injury severity. A resampling technique was used to deal with the problem of bias. SHapley Additive exPlanations (SHAP) analysis interpreted and ranked the factors contributing to crash injury. Two forms of modeling, namely multi and binary classification were adopted. Among the three models, XGBoost achieved the highest values of performance indicators including classification accuracy (71%), precision (70%), recall (71%), F1-scores (70%), and area under the Receiver Operating Characteristic (ROC) curve (AUC) (0.87) when used for multi-category classifications. While adopting the target as a binary classification, XGBoost again outperformed the other classifications with an accuracy of 94% and an AUC of 0.98. The SHAP results from both global and local interpretations illustrated that the accidents classified under property damage only, were primarily categorized by their consequences and the number of vehicles involved. The type of road and lighting conditions were mainly defined by the injury class. The death class was classified with respect to temporal parameters, including month and day of the week, as well as road type. Assessing the factors associated with the severe injuries caused by road traffic accidents will assist policy-makers in developing safety mitigation strategies in the Qassim Region, as well as in other regions of Saudi Arabia.
82 0