Traffic Fatality Prediction using Machine Learning Algorithms: Performance Analysis and Comparison Study

Many nations are concerned about the high number of traffic fatalities, which has led to a search for efficient prevention methods. This study addresses road fatalities brought on by drunk driving incidents. The study uses machine learning algorithms to detect and comprehend patterns related to fatalities brought on by drunk driving. The study builds prediction models that can help with early detection and policy development using several variables, including crash features, demographic information, and historical trends. Datasets on traffic fatalities are analyzed using a variety of supervised machine learning algorithms, including Random Forests (RF), Decision Tree (DT), Naive Bayes (NB), Logistic Regression (LR), and Support Vector Machine (SVM). According to the results of this study, the RF model has demonstrated great potential as a reliable predictor of fatalities caused by drunk driving. RF algorithm achieves better accuracy than other methods.