摘要
针对传统的模糊C均值聚类(fuzzy C-means clustering)算法容易陷入局部最优解,并且对初始值敏感的缺陷,提出一种基于改进的遗传算法的模糊聚类算法。该算法针对遗传算法的早熟问题提出一种改进的遗传算法,并将其应用于FCM算法,来寻找全局最优的聚类中心。实验表明,该算法与基于传统遗传算法的FCM算法相比,具有更强的寻优能力,更优的聚类效果。
The traditional fuzzy C-means（ FCM） clustering algorithm is prone to fall into the solution of local optimum and is sensitive to initial value. Aiming at these drawbacks,a fuzzy C-means based on the improved genetic algorithm is presented. The improved genetic algorithm is employed to optimise the FCM algorithm,finding the cluster center of the global optimum. Finally,the experimental results show that compared with the traditional FCM,the proposed algorithm has stronger optimisation ability and better clustering effect
出处
《智能系统学报》
CSCD
北大核心
2015年第4期627-635,共9页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(61172144)
国家科技支撑计划资助项目(2013BAH12F02)
辽宁省教育厅科学研究一般资助项目(L201432)
关键词
模糊C均值算法
聚类分析
遗传算法
动态分析
模糊聚类
初始值
避免早熟
全局最优
局部最优
fuzzy C-means clustering
cluster analysis
genetic algorithm
dynamic analysis
fuzzy clustering
initial values
premature contraction avoidance
global optimum
local optimum
