A Driving Behavior Distribution Fitting Method Based on Two-Stage Hybrid User Classification

Determining the distribution fitting of traditional private vehicle user driving behavior is an effective way to understand differences between different users and provides valuable information on travel demands. The classification significant product improvement, precision marketing, recommendations. This study proposed a method which includes four aspects: (1) data collection; (2) preprocessing; (3) analysis—a two-stage hybrid classification, (4) method. A used cluster users. First, first-stage extracts daily typical time–mileage-series patterns (TMTP) obtain time characteristics. also mean standard deviation mileage traveled (DVMT) express Next, are divided by K-means based characteristics demands from first stage. Finally, three-parameter log-normal fit DVMT types. Comparison with clustering proportion each vehicle’s series in TMTP types, this reveals that new methods provide advantages analyzing high reference value for enterprises making electric range recommendations, car market segmentation, policy decisions.