Predictive Model for Accident Severity

In the development of sustainable transportation, traffic safety is a significant matter, and predicting the severity of traffic accidents is still a critical problem in the traffic safety field. However, the utilized road traffic accidents (RTAs) datasets suffer from imbalance distribution. This problem leads to a decrease in classification performance, specifically in predicting the minority classes. This paper aims to treat the class imbalance problem through Synthetic Minority Over-sampling Technique (SMOTE), Support Vector Machine-SMOTE (SVMSMOTE), Borderline-SMOTE(BL-SMOTE), and Adaptive Synthetic (ADASYN), along with proposing an accurate predictive model for accident severity. Different accident severity models are employed, namely Random Forest (RF), KNearest Neighbor (KNN), Naïve Bayes (NB) classifiers, Decision Tree (DT), and Extra Trees (ET). These models are tested using real-world datasets. Several evaluation metrics are used to evaluate the proposed model. Experimental results show that the proposed model significantly improves predicting both the minority and majority classes. These results indicate the robustness and reliability of the proposed predictive model in enhancing road traffic safety and management.