A Segment-Based Trajectory Similarity Calculation Method SDTW

The analysis based on spatio-temporal data has become a hot topic in the field of machine learning. Urban traffic trajectory clustering is one of the key points of urban traffic data mining, and trajectory similarity calculation is the basis of traffic trajectory clustering. In the light of the weakness of current mainstream similarity calculation method under complex roads that proposed one segmentation-based trajectory similarity calculation method(SDTW*). The algorithm fully considers the shape of the trajectory and has good performance. The experimental part uses different similarity algorithms, and combines the hierarchical clustering algorithm to cluster the actual vehicle trajectory, and selects the average contour coefficient and the cluster success rate as the evaluation indicators. The results show that the average contour coefficient of the proposed algorithm is 33.86% and 12.94% higher than that of DTW and SDTW, respectively, and the clustering success rate is also improved to some extent.