Combining K-means and particle swarm optimization for dynamic data clustering problems
Yu‐Cheng Kao, Szu-Yuan Lee · 2009
This paper presents a new dynamic data clustering algorithm based on K-means and combinatorial particle swarm optimization, called KCPSO. Unlike the traditional K-means method, KCPSO does not need a specific number of clusters given before performing the clustering process and is able to find the optimal number of clusters during the clustering process. In each iteration of KCPSO, a discrete PSO is used to optimize the number of clusters with which the K-means is used to find the best clustering result. KCPSO has been developed into a software system and evaluated by testing some datasets. Encouraging results show that KCPSO is an effective algorithm for solving dynamic clustering problems.