A novel hybrid clustering algorithms with chaotic particle swarm optimization

作者：Zhang J.^*; Wang Y.; Feng J.

来源：Journal of Computational Information Systems, 2012, 8(21): 8827-8834.

• Chaotic
• Clustering
• Dynamically
• Hybrid
• K-means
• Perturbation
• PSO

In order to overcome the premature convergence in the particle swarm optimization algorithm, dynamically chaotic perturbation is introduced to form a dynamically chaotic PSO, briefly denoted as DCPSO. To get rid of the drawbacks of simply finding the convex cluster and being sensitive to the initial partitions in k-means algorithm, a novel hybrid clustering algorithm combined with the DCPSO is proposed. The difference between the work and the existing ones is that DCPSO is firstly introduced into k-means. From the experimental results made on several data sets, we can see that the proposed clustering algorithms can get completely rid of the shortcomings in k-means algorithms.

• 出版日期2012
• 单位西安电子科技大学