Enhancing Thailand Road Safety through Personal Driving Style Classification Using Machine Learning Model

In Thailand, road safety remains a critical concern due to various factors, including diverse driving styles. To address this issue, this study proposes an approach utilizing machine learning models to classify personal driving styles, aiming to enhance road safety measures. By analyzing driving behavior data collected from Inertial Measurement Unit sensors (IMU), which is a telematics device. This study presents the simulation of a driver behavior dataset to determine the appropriate machine learning model. Then analyzes the dataset, selects features suitable for feature scaling, and trains models using machine learning algorithms, including Support Vector Machine (SVM), Logistic Regression (LR), Random Forest (RF), and Decision Tree (DT). Our results reveal that the Random Forest algorithm outperforms others, achieving an accuracy of 89%, precision of 78%, recall of 89%, and an f1-score of 82%. This highlights the effectiveness of the proposed approach in accurately driver behavior, which is classified as safety driving and unsafety driving, the study lays the foundation for the development of tailored road safety initiatives aimed at reducing accidents and promoting responsible driving practices in Thailand.