一种改进的模糊C-均值(FCM)聚类算法  

A modified fuzzy C-means(MFCM) clustering algorithm

作　　者：安良[1] 胡勇[1] 胡良梅[1] 孟玲玲[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2003年第3期354-358,共5页Journal of Hefei University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(60175011);安徽省自然科学基金资助项目(01042301);教育部优秀青年教师资助项目

摘　　要：模糊C-均值(FCM)聚类算法受初始化影响较大,在迭代时容易陷入局部极小,鉴于遗传算法(GA)的并行全局搜索能力,文章将遗传算法引入进来对FCM聚类算法加以改进,并对所提出的新算法与经典算法的迭代步数和运行时间进行比较。实验结果表明:该算法与FCM聚类算法相比收敛速度更快,迭代步数更少。The fuzzy Cmeans(FCM) clustering algorithm is sensitive to the situation of the initialization and easy to fall into the local minimum when iterating. In view of the parallel and whole search capacity of the genetic algorithm(GA),the GA is introduced to improve the fuzzy Cmeans clustering algorithm. The modified clustering algorithm is compared with the classical clustering algorithm by testing the iteration steps and the CPU time of the two algorithms. Experimental results show that the new algorithm is faster than the FCM algorithm and the iterating times are less .

关 键 词：改进模糊C-均值聚类算法 FCM聚类算法 遗传算法 迭代步数 运行时间 

分 类 号：O159[理学—数学] TP18[理学—基础数学]