2015 International Symposium of Computer Application and Information Technology (IsCait 2015)

210 / 1971-01-01 00:00:00

Entropy-based K-means Algorithm Combined With Particle Swarm Optimization

5534,931,5535

To deal with the problems of being sensitive in choosing cluster centers and easy convergence, an algorithm based on improved K-means cluster algorithm combined with Particle Swarm Optimization algorithm is presented. In this paper, the group fitness variance is adopted in order to decide when to execute K-means clustering. Meanwhile, weighted Euclidean distance is introduced into the process of clustering to improve the stability. Experimental results show that the new algorithm has good clustering stability and better global convergence.