摘要
属性约减是粗糙集理论的重要研究内容之一。由于Z.Pawlak经典粗糙集模型在处理集合间隶属关系过于简单的缺陷,文章提出了以集合间距离作为集合隶属关系的判别依据,对属性依赖度和重要度重新进行了定义,从而对属性约减算法进行改进。最后,通过一个数据模型的验证,改进后的算法能够更有效地滤除冗余属性,保留关键属性。
Knowledge reduction is one of the most important issues in rough sets theory.Because of the defects that Z.Pawlak rough set model in dealing with the collection of affiliation is too simple,this paper put forward to use collection＇s distance to discriminate collection,then redefine the dependence and importance of attribute,so improve the attribute reduction algorithm.Finally,under the validation of a data model,the improved algorithm can effectively filter out redundant properties,retain key attributes.
出处
《计算机与数字工程》
2010年第11期48-51,共4页
Computer & Digital Engineering
关键词
粗糙集
数据挖掘
属性约减
聚类分析
距离
yough set
data mining
attribute reduction
cluster analysis
distance
