Zusammenfassung
Ubiquitous computing has also reached the insurance indus-try in the form of Usage Based Insurance models. Modern rates use sensordata to offer the user suitable pricing models adapted to his character.Our overview shows that insurance companies generally rely on drivingbehaviour to assess a user in risk categories. Based on the collected data,a new attack using kNN-DTW shows that, with ...
Zusammenfassung
Ubiquitous computing has also reached the insurance indus-try in the form of Usage Based Insurance models. Modern rates use sensordata to offer the user suitable pricing models adapted to his character.Our overview shows that insurance companies generally rely on drivingbehaviour to assess a user in risk categories. Based on the collected data,a new attack using kNN-DTW shows that, with the derived informa-tion, the identification of a driver in a group of all users of a vehicle ispossible with more than 90% accuracy and therefore may represent amisuse of the data collection. Thus, motivated by the General Data Pro-tection Regulation, questions regarding anonymisation become relevant.The suitability of standard methods known from Big Data is evaluatedin the complex scenario Pay-How-You-Drive using both real-world andsynthetic data. It shows that there are open questions considering thefield of privacy-friendly Pay-How-You-Drive models.