Analysis on Zone-based Movement Pattern Using Taxi Trajectory Data

We propose an integrated approach to discover both zones and movement trajectories among zones, which referred to as zone-based movement pattern (ZMP), from taxi trajectory data. This method discovers the zones by merging ZMPs, which keeps the directionality of movement, thematic attributes and distance relationship of zones by the adjacent constraints consists of distant and thematic attributes. By joint average frequencies, we can identify new ZMP by iteratively calculating the best candidate ZMPs to be merged then. In addition, evaluation measures of ZMP are suggested in terms of factors such as coverage, accuracy and a tradeoff of both them. The effectiveness of the proposed approach is demonstrated through a real-world data set obtained, the experiment result shows that the approach can merge the existing zones to discover new ZMP rationally.