Improved Clustering Algorithm Based on Hypercube
Yingying Xia, Liang Zhou · 2022
The hypercube clustering algorithm based on segmented clustering framework has a good effect on trajectory clustering, but when the trajectory points deviate greatly, the direction of the hypercube obtained by the hypercube clustering algorithm will have a large error, which affects the clustering effect. Therefore, an improved clustering algorithm based on hypercube is proposed. The improved algorithm segmented the trajectories by introducing the motion direction, transformed each trajectory into a hypercube sequence, expanded the checking dimension of common sub-trajectories, and gave the intersecting checking and clustering algorithm of common sub-trajectories. In this paper, three real data sets are used to carry out experiments. The results show that the improved algorithm has better clustering effect, and the average clustering number and average clustering accuracy are improved by 65.81 percent and 15.77 percent respectively.