Driver Risk Assessment Using Traffic Violation and Accident Data by Machine Learning Approaches

Abstract

Along with high speed urbanization and motorization, road traffic accidents have become a severe problem in China. Drivers' operation error and risk-taking behavior is a leading cause of traffic accidents. Under this condition, the demand of drivers' traffic safety assessment keeps increasing, especially for professional drivers like passenger drivers and freight drivers. This work proposes a data mining framework of drivers' traffic safety assessment using drivers' personal information, traffic violations and traffic accident data records. Model validation and result interpretation are given, showing the rationality and usefulness of our proposed approach.