基于Rough集的连续属性离散化方法  

Discrete method of continuous attributes based on rough set

作　　者：尹旭日[1] 

机构地区：[1]解放军汽车管理学院基础部,安徽蚌埠233011

出　　处：《计算机工程与设计》2006年第11期2038-2040,共3页Computer Engineering and Design

摘　　要：连续属性离散化是Rough集理论应用中面临的主要问题之一。提出了一种基于的Rough集连续属性离散化方法。首先提出主泛化决策等概念,在数据过滤方法的基础上,利用等价类的合并对属性离散化。实验表明,利用该方法对数据进行离散预处理后提取的规则具有较好的分类预测准确性。The discretization of continuous attributes is one oft.he most main problems in applying of rough set theory. A new method ofdiscretization based on rough set theory is presented. The concept of generalized decision is introduced. Its basic idea is to merge the equivalence classes of （Q on basis of the methods of data filtering. Experimental results show this algorithm is efficient, comprehensible and suitable for a pre-processing step before a data-mining method is employed.

关 键 词：离散化 ROUGH集 泛化决策 等价类 置信度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]