作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2008, Vol. 34 ›› Issue (20): 200-202. doi: 10.3969/j.issn.1000-3428.2008.20.073

• 人工智能及识别技术 • 上一篇    下一篇

基于遗传算法的K均值聚类分析

赖玉霞1,刘建平1,杨国兴2   

  1. (1. 浙江理工大学信息电子学院,杭州 310018;2. 浙江天健会计师事务所,杭州 310012)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-10-20 发布日期:2008-10-20

K-Means Clustering Analysis Based on Genetic Algorithm

LAI Yu-xia1, LIU Jian-ping1, YANG Guo-xing2   

  1. (1. College of Information and Electronics, Zhejiang Science and Technology University, Hangzhou 310018;2. Zhejiang PanChina Accounting Firm, Hangzhou 310012)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-10-20 Published:2008-10-20

摘要: 传统K均值算法对初始聚类中心敏感,聚类结果随不同的初始输入而波动,容易陷入局部最优值。针对上述问题,该文提出一种基于遗传算法的K均值聚类算法,将K均值算法的局部寻优能力与遗传算法的全局寻优能力相结合,在自适应交叉概率和变异概率的遗传算法中引入K均值操作,以克服传统K均值算法的局部性和对初始中心的敏感性,实验证明,该算法有较好的全局收敛性,聚类效果更好。

关键词: K均值算法, 聚类中心, 遗传算法

Abstract: Traditional K-Means algorithm is sensitive to the initial centers and easy to get stuck at locally optimal value. To solve such problems, this paper presents an improved K-Means algorithm based on genetic algorithm. It combines the locally searching capability of the K-Means with the global optimization capability of genetic algorithm, and introduces the K-Means operation into the genetic algorithm of adaptive crossover probability and adaptive mutation probability, which overcomes the sensitivity to the initial start centers and locality of K-Means. Experimental results demonstrate that the algorithm has greater global searching capability and can get better clustering.

Key words: K-Means algorithm, clustering center, genetic algorithm

中图分类号: