摘要
简要介绍了系统聚类法的各种常用方法 ,对 Ward方法进行了分析 ,说明爬山法一般得不到全局最优分类 ,因此提出结合 Ward方法 ,用遗传算法进行聚类 ,特别是对交叉操作进行了设计 ,从而将遗传算法用于聚类分析 .最后的实例表明 ,遗传算法在处理多样本、多属性、多类别的问题时 。
The various kinds of methods in common use in hierachical clustering are introduced.The Ward method is analyzed.As mountain climbing method is not able to obtain the best overall classification, it combines genetic algorithm with Ward method for hierachical clustering.The cross operator is designed in particular to adapt to clustering analysis.The example given shows that the genetic algorithm is efficient in dealing with problems of multi samples,multi properties and multi classifications.
出处
《武汉交通科技大学学报》
2000年第3期301-304,共4页
Journal of Wuhan University of Technology(Transportation Science & Engineering)
关键词
系统聚类
Ward法
爬山法
遗传算法
hierachical clustering
Ward method
mountain climbing method
genetic algorithm
