Driving Style Identification with Unsupervised Learning
One way to optimise insurance prices and policies is to collect and to analyse driving trajectories: sequences of 2D-points, where time distance between any two consequitive points is a constant. Suppose that most of the drivers have safe driving style wi
- PDF / 1,440,406 Bytes
- 15 Pages / 439.37 x 666.142 pts Page_size
- 60 Downloads / 210 Views
Abstract. One way to optimise insurance prices and policies is to collect and to analyse driving trajectories: sequences of 2D-points, where time distance between any two consequitive points is a constant. Suppose that most of the drivers have safe driving style with similar statistical characteristics. Using above assumption as a main ground, we shall go through the list of all drivers (available in the database) assuming that the current driver is “bad”. We shall add to the training database several randomly selected drivers assuming that they are “good”. By comparing the current driver with a few randomly selected “good” drivers, we estimate the probability that the current driver is bad (or has significant deviations from usual statistical characteristics). Note as a distinguished particular feature of the presented method: it does not require availability of the training labels. The database includes 2736 drivers with 200 variable length driving trajectories each. We tested our model (with competitive results) online during Kaggle-based AXA Drivers Telematics Challenge in 2015. Keywords: Motion mining · Unsupervised learning · Similarity measures · Learning of similarity · Big data · Sequential data
1
Introduction
From the point of view of optimizing of insurance pricing accuracy, it is desirable to be able to identify a driver by his driving style. Using inertial sensor data we can analyse and understand driving style. Consequently, such analysis can help reduce dangerous driving [1]. In the world today, there are over one billion cars with different drivers interacting with each other on the roads. Each driver has their own driving style, which could impact safety, fuel economy, and road congestion, among many things. The precise relationships between driving style and their effects have not been well characterized, although there is some general consensus that “aggressive” driving (e.g. speeding, hard braking or sharp turning) has a mostly negative impact [2]. In the traditional system, a consumer calls an insurance company, provide some basic information, and get a quote for auto insurance based on type of car, c Springer International Publishing Switzerland 2016 P. Perner (Ed.): MLDM 2016, LNAI 9729, pp. 155–169, 2016. DOI: 10.1007/978-3-319-41920-6 12
156
V. Nikulin
age, gender, marital status, location, driving history, and credit history of the customer. These attributes act as proxies for your auto risk, and the riskier you are the more likely you are to file a claim. These attributes allow an insurance firm to stack you up against the rest of population and see where you are likely to fall in terms of risk. They are playing a guessing game based on averages. In the last years, there have been active research toward developing systems that make driving safer. In the current insurance markets, consumers have rejected the so-called Pay-As-You-Drive due to two main reasons: the required installation of “black-boxes” in vehicles makes drivers perceive the monitoring as intrusive, and the installation
Data Loading...
