摘要
针对模糊K-Means算法随机选择初始数据中心而导致的聚类效果不稳定的问题,提出了一种粒子群优化的模糊K-Means改进聚类算法。首先,定义了一个确定聚类数K和初始数据中心的算法,然后将算法得到的初始数据中心作为初始粒子,采用粒子群优化算法进行寻优获得最优数据中心,最后再使用模糊K-Means算法根据最优数据中心进行聚类。在UCI数据集上的实验结果表明文中算法能准确地实现分类,具有较强的全局寻优能力、较少的寻优时间和较快的收敛能力,能有效地解决目标分类问题。
In order to overcome the defect of the unsteady clustering effect caused by the random choosing initial data center of fuzzy K --Means algorism, a clustering algorism is proposed based on particle swarm optimization and fuzzy K--Means algorism. Firstly, a algorism defined for obtaining clustering count K and initial data center is proposed, then the initial data center is used as the initial particle, and the particle swarm algorism is used to get the optimum data center, finally the fuzzy K--Means algorism is clustering according to the optimum data center. The experiment on UCI data set shows the method in this paper can accurately realize classification with the strong global optimization ability, the less time and rapid convergence ability, it can effectively solve the goal classification problem.
出处
《计算机测量与控制》
北大核心
2013年第5期1266-1268,共3页
Computer Measurement &Control
关键词
粒子群
模糊
分类
K均值
聚类
particle swarm
fuzzy
classification
K--Means
clustering
