摘要
为了改进K-means聚类算法的不足,把混合粒子群优化算法引入到K-means聚类算法中,重新选取编码方式并构造适应度函数,在此基础上提出了一种改进的K-means聚类算法;通过两个经典数据集的测试,实验结果表明:改进的算法比K-means算法具有更好的全局寻优能力、更快的收敛速度,且其解的精度更高对初始聚类中心的敏感度降低。
This paper incorporates hybrid particle swarm optimization algorithm into the K -means to overcome the local search of K - means algorithm, and adds the penalty function to reconstruct the fitness function, and proposes an improved K -means Cluster Algorithm, the computational experimental results on two benchmark dataset have shown that the improved K- means has better globe search capability, faster convergence velocity and is to attain higher precision value than K- means algorithm.
出处
《重庆工商大学学报（自然科学版）》
2009年第2期144-147,共4页
Journal of Chongqing Technology and Business University:Natural Science Edition
基金
重庆市科委自然科学基金计划资助项目(CSTC.2007BB2372)
关键词
混合粒子群优化算法
K-均值
聚类算法
hybrid particle swarm optimization algorithm
K - means
cluster algorithm
