Mohammad Mehdi Sepehri - Assosiate professor Tarbiat Modares University
Nasim Ghanbar Tehrani - PhD Candidates Tarbiat Modares University
Hoda Davarzani - PhD Candidates Tarbiat Modares University
Arezoo Atighehchian - PhD Candidates Tarbiat Modares University

This research applies data mining techniques to discover the relationship between motor vehicle accidents and the comprehensive information about people, vehicles, and conditions recorded in Police Accident Reports. This is one of the few research papers that address the data mining techniques to explore the motor vehicle crashes and related factors. The data used in this research are obtained from the traffic police’s accident database, focused on the third traffic zone in Tehran from years 2004 to 2005. The data are first clustered using a suitable clustering algorithm such as the K-mean method. Then, the patterns and rules of the data are explored and using these descriptions, some refinements in behaviors are proposed. The refinements in three levels for citizens, urban police and top level policy makers are put forward based on the results. Some advices, based on the results, are proposed for consideration of each level, where following these advices will decrease accidents and increase safety. By using this innovative method, regardless of the constraints of this study, useful knowledge about the high level of safety and driving and accident management will be explored and tailored to the needs of the three levels.

Urban Transportation, Data Mining, Incident Management, safety