面向属性的RST在数据挖掘中的应用  

Attribute Oriented Rough Sets Theory and Its Applications in Data Tapping

作　　者：朱定华[1] 魏媛媛 魏长华[3] 

机构地区：[1]华中理工大学电子与信息工程系,430074 [2]华中师范大学电子与计算机研究所,430074 [3]华中师范大学计算机科学系,430074

出　　处：《华中理工大学学报》2000年第2期80-83,共4页Journal of Huazhong University of Science and Technology

基　　金：湖北省自然科学基金资助项目!( 98J76)

摘　　要：基于对 RST的基本概念和近似空间的形式描述 ,定义了不确定信息系统 ,并引入面向对象属性的泛化和约简操作算法 .在此基础上 ,提出了一个面向属性的 RST的数据挖掘方法 ,即基于系统的关系数据库形成差别矩阵 ,并对差别矩阵施行最佳约简和核操作 ,给出在逻辑意义下的决策规则 ,从而完成领域的知识发现 .Reasoning with uncertain information is a key problem when dealing with information about the real world. An uncertain information system (UIS) is defined with the basic concept of RST including the rough sets theory, upper approximation, lower approximation and boundary, as well as the formal description of approach space. The undistinguishable relationship, the attribute oriented generalization and the reduced operation algorithm are discussed. Then a tapping method for attribute oriented RST data is presented. Applying the best reduction and kernel operation to the differential matrix, the decision rules in logical meaning are given. The application of the theoretical treatment developed for medical diagnosis is described.

关 键 词：粗集理论 数据挖掘 不确定信息系统 面向属性 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]