Classification of scale-sensitive telematic observables for riskindividual pricing

Using modern telematics technologies for car insurance, it is no particular challenge to produce an intractably large amount of kinematic and contextual information about driving profiles of motor vehicles. In order to evaluate this data with respect to both efficient and effective use in scoring and subsequent actuarial pricing, we propose a scale-sensitive approach that treats observations on semantically different levels. Furthermore we discuss the application of methods necessary to assess the information of different scale levels including signal processing, pattern recognition and Fourier analysis. In this way we show how maneuvers, trips and trip sections as well as the total insurance period can be analyzed to individually or collectively gain significantly scoreable insights into individual driving behaviour.