A Comparison Between Alignment and Integral Based Kernels for Vessel Trajectories

In this paper we present a comparison between two important types of similarity measures for moving object trajectories for machine learning from vessel movement data. These similarities are compared in the tasks of clustering, classication and outlier detection. The rst similarity type are alignment measures, such as dynamic time warping and edit distance. The second type are based on the integral over time between two trajectories. Following earlier work we dene these measures in
the context of kernel methods, which provide state-of-the-art, robust algorithms for the tasks studied. Furthermore, we include the in uence of applying piecewise linear segmentation as
pre-processing to the vessel trajectories when computing alignment measures, since this has been shown to give a positive eect in computation time and performance.
In our experiments the alignment based measures show the best performance. Regular versions of edit distance give the best performance in clustering and classication, whereas the softmax variant of dynamic time warping works best in outlier detection. Moreover, piecewise linear segmentation has a positive eect on alignments, which seems to be due to the fact salient
points in a trajectory, especially important in clustering and outlier detection, are highlighted by the segmentation and have a large in uence in the alignments.