Ship Trajectory Clustering Based on Affinity Propagation and Spectral Clustering

Apply

Ship Trajectory Clustering Based on Affinity Propagation and Spectral Clustering

Journal Title: 河南科技大学学报（自然科学版） - Year 2018, Vol 39, Issue 1

Abstract

In order to solve the problem that the common K-means algorithm of spectral clustering for characterized vector spaces was very sensitive to the initial value, a spectral clustering algorithm based on affinity propagation (AP) was proposed.Firstly, the dynamic time warping (DTW) was used to measure the structure similarity between trajectories automatic identification systems (AIS) , and the distance matrix was acquired.Secondly, the traditional spectral clustering algorithm was improved by fast AP clustering algorithm, and the inland bridge area AIS trajectory data was verified by example based on the designated category numbers.The numerical experiment results show that the fusion algorithm has better robustness compared with conventional spectral clustering algorithm on the basis of invariable time complexity.The accuracy of the experiment is increased by 5.24%.

Authors and Affiliations

Yuxiao LI, Chuansheng WU, Wen LIU, Huanhuan LI

Keywords

trajectory clustering; dynamic time warping distance; spectral clustering; similarity measure; affinity propagation clustering;

EP ID EP464631
DOI -
Views 87
Downloads 0