Volume XLII-4/W18
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 737–740, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-737-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 737–740, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-737-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

  18 Oct 2019

18 Oct 2019

A COMPARISON OF EFFICIENCY OF THE OPTIMIZATION APPROACH FOR CLUSTERING OF TRAJECTORIES

A. Moayedi1, R. A. Abbaspour1, and A. Chehreghan2 A. Moayedi et al.
  • 1School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran, Iran
  • 2Mining Engineering Faculty, Sahand University of Technology, Tabriz, Iran

Keywords: Trajectories, Clustering, K-means, Optimization, Meta-heuristic Algorithms, DTW

Abstract. Clustering is an unsupervised learning method that used to discover hidden patterns in large sets of data. Huge data volume and the multidimensionality of trajectories have made their clustering a more challenging task. K-means is a widely used clustering algorithm applied in the trajectory computation field. However, the critical issue with this algorithm is its dependency on the initial values and getting stuck in the local minimum. Meta-heuristic algorithms with the goal of minimizing the cost function of the K-means algorithm can be utilized to address this problem. In this paper, after suggesting a cost function, we compare clustering performance of seven known metaheuristic population-based algorithms including, Grey Wolf Optimizer (GWO), Particle Swarm Optimization (PSO), Sine Cosine Algorithm (SCA), and Whale Optimization Algorithm (WOA). The results obtained from the clustering of several data sets with class labels were assessed by internal and external clustering validation indices along with computation time factor. According to the results, PSO, and SCA algorithms show the best results in the clustering regarding the Purity, and computation time metrics, respectively.