Usage-based vehicle insurance: Driving style factors of accident probability and severity

This paper introduces an approach to telematics device data application in automotive insurance. We conduct a comparative analysis of different types devices that collect information on vehicle utilization and driving style its driver describe advantages disadvantages these devices. The possible formats are described methods their processing format convenient for modeling proposed. also introduce classify the accidents’ strength. Using all available information, we estimate accident probability models accidents identify optimal set factors each models. assess quality resulting using both in-sample out-of-sample estimates.