GPS Trajectory Feature Extraction for Driver Risk Profiling

Faculty/Professorship: Information Systems and Energy Efficient Systems  
Author(s): Paefgen, Johannes; Michahelles, Florian; Staake, Thorsten
Title of the compilation: Proceedings of the 2011 international workshop on Trajectory data mining and analysis
Corporate Body: Ubicomp '11: The 2011 ACM Conference on Ubiquitous Computing Beijing, China — September 17 - 21, 2011
Publisher Information: New York, NY, USA : ACM
Year of publication: 2011
Pages: 53-56
Language(s): English
DOI: 10.1145/2030080.2030091
In this paper we develop a method and relevant feature constructs for the measurement of accident risk exposure from a large sample of real-world GPS data that includes accident and accident-free drivers. For trip frequency and accumulated driven distance features, an evaluation of their discriminatory power is given based on computational results. In our conclusion, we briefly discuss suitable classification approaches and limitations arising from external validity considerations.
Keywords: Usage-based Insurance, Feature Extraction, Accident Risk, Spatio-temporal Trajectories, Driver Behavior
Type: Conferenceobject
Year of publication: 7. April 2014