Research on Trajectory Clustering Algorithm for Public Opinion Users
Download as PDF
DOI: 10.25236/meici.2019.074
Author(s)
Fuyu Lu, Yonglin Leng, and Xiaohong Sun
Corresponding Author
Fuyu Lu
Abstract
Clustering movement trajectories to get the motion feature of object is one of the goals of the trajectory clustering. Aiming at the large scale trajectory data, to address the low efficiency of clustering, this paper proposes a hierarchical trajectory clustering algorithm based on time series (HTCTS). The algorithm first divides trajectory data by time interval, and then respectively cluster the sub paths. Finally, for all cluster subset, HTCTS executes cluster algorithm again to produce the final clustering results. The experimental results show that HTCTS algorithm in clustering efficiency and quality is superior than the trajectory clustering algorithms which cluster the whole dataset directly.
Keywords
Trajectory clustering, AP algorithm, Public Opinion, Similarity