计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (2): 128-130.DOI: 10.3778/j.issn.1002-8331.2009.02.037

• 数据库、信号与信息处理 • 上一篇    下一篇

连续属性的频数监督断点离散化技术

林汀辉,史 亮,姜青山   

  1. 厦门大学 软件学院,福建 厦门 361005
  • 收稿日期:2008-07-08 修回日期:2008-11-17 出版日期:2009-01-11 发布日期:2009-01-11
  • 通讯作者: 林汀辉

Frequency-supervised-breakpoint based discretization algorithm

LIN Ting-hui,SHI Liang,JIANG Qing-shan   

  1. Software School,Xiamen University,Xiamen,Fujian 361005,China
  • Received:2008-07-08 Revised:2008-11-17 Online:2009-01-11 Published:2009-01-11
  • Contact: LIN Ting-hui

摘要: 提出了一种频数监督断点的离散化算法。该算法利用所提出的频数监督断点思想产生初始断点,并在此基础上进行断点简约。实验结果表明该算法所产生的断点不仅符合实际数据分布,而且更为合理、精练。

关键词: 频数监督断点, 不可分辨, 离散化

Abstract: In this paper,a discrete algorithm is proposed based on the frequency supervised breakpoint which is applied to the conditional continuous attributes.The algorithm adopts the idea of frequency supervised breakpoint that is brought forward to generate initial breakpoints.On the basis of the preparatory work,reduction of breaktpoints has been performed.The result obtained shows clearly the breakpoints generated by this algorithm not only are in line with the actual data distribution but also they are more reasonable,refined.

Key words: frequency-supervised-breakpoint, equivalent, discretization