计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (30): 148-149.DOI: 10.3778/j.issn.1002-8331.2008.30.045

• 数据库、信号与信息处理 • 上一篇    下一篇

决策表连续属性离散化的一种方法

王 柯,朱启兵,崔宝同   

  1. 江南大学 通信与控制工程学院,江苏 无锡 214122
  • 收稿日期:2007-12-03 修回日期:2008-02-01 出版日期:2008-10-21 发布日期:2008-10-21
  • 通讯作者: 王 柯

Method of discretization of continuous attributes of decision table

WANG Ke,ZHU Qi-bing,CUI Bao-tong   

  1. School of Communication and Control Engineering,Jiangnan University,Wuxi,Jiangsu 214122,China
  • Received:2007-12-03 Revised:2008-02-01 Online:2008-10-21 Published:2008-10-21
  • Contact: WANG Ke

摘要: 提出了一种基于区间数据分布特征的决策表连续属性离散化的方法。方法在断点的选择上考虑了属性值的出现频率,在区间内的一致性和区间之间的差异性基础上,利用条件信息量作为反馈信息合并区间。通过实验分析表明了算法的有效性,能保持决策表较高的分类能力,提高约简效率。

关键词: 决策表, 连续属性, 信息量, 离散化

Abstract: The paper puts forward a method of discretization of continuous properties based on distribution characterization of interval data.It considers the frequency of attribute values in choice the cut point,uses information quantity as feedback on the basis of consistency within the interval and differences between the intervals,merger the intervals.The experiment result shows that the algorithm is effective.It can improve the efficiency of knowledge reduction when the decision table keeps stable.

Key words: decision table, continuous attributes, information quantity, discretization