计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (14): 128-130.

• 数据库、信号与信息处理 • 上一篇    下一篇

一种改进的分解-合并聚类方法

安世全,丁进标,高 涛   

  1. 重庆邮电大学 计算机科学与技术学院,重庆 400065
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-05-11 发布日期:2011-05-11

Improved separate-and-amalgamate clustering methodology

AN Shiquan,DING Jinbiao,GAO Tao   

  1. College of Computer Science and Technology,Chongqing University of Posts and Telecommunications,Chongqing 400065,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-05-11 Published:2011-05-11

摘要: 为了对一个对象集合进行聚类,提出一种基于分解-合并方法的聚类算法,方法综合了自顶向下的分解过程和自底向上的合并过程。在分解过程中,使用分割算法建立一个树,树的叶子就是对象集合的元素;合并过程中能够快速发现树的最优部分。实验结果表明该算法在聚类效率和性能上优于传统算法。

关键词: 聚类, 数据挖掘, 信息检索

Abstract: Present a separate-and-amalgamate methodology for clustering a set of objects that combines a top-down “separate” phase with a bottom-up “amalgamate” phase.For the separate phase,which produces a tree whose leaves are the elements of the set,an efficient separate algorithm is suggested;The amalgamate phase quickly ?nds the optimal partition that respects the tree.The experiment proves that this algorithm can get much better performance than traditional algorithms.

Key words: clustering, data mining, information retrieval