Swarm Clustering Algorithm: Let the Particles Fly for a while
Wenjie Zhu, Wenjian Luo, Ni Li, Nannan Lu · 2018
Swarm Intelligence (SI) and Evolutionary Algorithms (EAs) have been widely used for cluster analysis of spatial data. However, in most existing SI, particles are encoded to represent the centers of clusters. In this paper, inspired by Particle Swarm optimization (PSO), a novel Swarm Clustering Algorithm (SCA) is proposed, which has the potential ability to deal with the data of the arbitrary number, shape and size of clusters. In SCA, a particle is a point in the dataset under cluster analysis. Thus, the number of particles in the swarm is equal to the size of the dataset. All particles interact dynamically with similar particles, and fly to the denser areas to form clusters. The experimental results show that our algorithm is effective.