摘要
利用粒子群优化(PSO)算法全局寻优的特点,很大程度上避免了模糊C-均值聚类(FCM)算法对初值敏感、易陷入局部收敛的缺陷。利用收敛速度快的K均值聚类法得到的聚类中心作为PSO算法初始聚类中心的参考,提出一种新的模糊C-均值聚类算法Improved PSO FCM。实验结果表明,论文算法提高了FCM的搜索能力,聚类更为准确,效率更高。
The fuzzy C-means clustering algorithm has the problems of local optimal value and sensitivity to initial values, which are overcomed by particle swarm optimization algorithm with the global optimization. A new fuzzy C-means clustering algorithm, Improved PSO FCM is proposed with the clustering centers obtained by K-means algorithm as the reference of the searching scope of PSO algorithm. The experimental results show the algorithm improves the searching capacity of FCM, and is more accurate and efficient.
出处
《计算机与数字工程》
2014年第9期1610-1612,1724,共4页
Computer & Digital Engineering