基于改进的粒子群优化的模糊C-均值聚类算法  

Fuzzy C-means Clustering Algorithm Based on Improved Particle Swarm Optimization

作　　者：王杨[1] 

机构地区：[1]辽宁石油化工大学计算机与通信工程学院,抚顺113001

出　　处：《计算机与数字工程》2014年第9期1610-1612,1724,共4页Computer & Digital Engineering

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.

关 键 词：粒子群优化算法 模糊C-均值聚类算法 初始聚类中心 K均值算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]