A Trajectory Data Clustering Method Based On Dynamic Grid Density

Abstract

Under the traditional method of frequent trajectory mining, the location of data is obtained through the GPS device. However, limited equipment accuracy may incur location ambiguity. In this paper, we propose a new trajectory data clustering method based on dynamic grid density, in order to remove this ambiguity. In this method, the trajectory space of an object is firstly divided into equal-sized squares dynamically. Then the trajectories of object are mapped to their corresponding square. Next, the density of each grid is calculated and all the frequent squares are acquired given the minimum support threshold. Lastly, the frequent area is obtained by merging the frequent squares acquired previously, using the boundary function provided. The experimental results show that this method provides an optional way of finding the frequent movement sequence.