Multidimensional trajectory mining and its application to medicine

This paper focuses on such a nature of human movements as a trajectory in two or three dimensional spaces and proposes a method for grouping trajectories as two-dimensional time-series data, consisting of the following two steps. Firstly, it compared two trajectories based on their structural similarity, determines the best correspondence of partial trajectories and calculates the dissimilarity between the sequences. Then clustering method are applied by using the dissimilarity matrix. Experimental results shows that this method succeeded in capturing the structural similarity between trajectories.