Travel mode classification of intercity trips using cellular network data

Many applications in transport planning require an understanding of travel patterns separated by mode. To use cellular network data as observations human mobility these applications, classification mode is needed. Existing methods for GPS-trajectories are often inefficient data, which has lower resolution space and time than GPS data. In this study, we compare three geometry-based supervised to classify trips extracted from intercity origin-destination pairs either road or train. understand the difficulty problem, a labeled dataset 255 two OD-pairs train evaluate performance. For OD-pair where routes not more four kilometers, 4.5% - 7.1% wrong, while can all correctly. Using large-scale 29037 trips, find that separation between classes less evident show choice impacts aggregated modal split estimate.