Driving Style Analysis Using Data Mining Techniques

This paper investigates the modeling of the personal driving style of various vehicle drivers based on several driving parameters. The purpose of such an endeavor is to classify the drivers according to their risk-proneness within the larger context of increasing traffic safety, which is a major concern worldwide. This information is valuable especially for those involved in fleet management and it can be used to improve and to make safer the driving style of various individuals who serve within that fleet. Equally important, such information could help any driver to see the danger within his or her driving style. Cluster and principal component analyses from exploratory statistics have been used to identify and explain drivers grouping according to their driving behavior. The driving parameters (behavioral indices) are collected from urban traffic by an in-house developed GPS-based device and sent to a data server for analysis.