Selective Regenerated Particle Swarm Optimization for Data Clustering

This paper proposes an improved particle swarm optimization (PSO) for data clustering. In order to increase the efficiency, suggestions on parameter settings is made and a mechanism is designed to prevent particles fall into the local optimal. To evaluate its effectiveness and efficiency, this approach is applied to data clustering problem. The proposed methods were tested on nine datasets, and their performance is compared with those of PSO, K-means and four other clustering methods. Results show that our schemes are both robust and suitable for solving data clustering problems.