改进的相容矩阵的属性约简算法  

Improved tolerance matrix attribute reduction algorithm

作　　者：阮慎[1,2] 徐章艳[1,3] 王炜[1] 杨炳儒[3] 

机构地区：[1]广西师范大学计算机科学与信息工程学院,广西桂林541004 [2]柳州师范高等专科学校数学与计算机科学系,广西柳州545004 [3]北京科技大学信息工程学院,北京100083

出　　处：《计算机工程与应用》2011年第32期49-50,147,共3页Computer Engineering and Applications

基　　金：国家自然科学基金No.60963008~~

摘　　要：在粗糙集理论中,知识约简已经成为其核心内容之一。针对相关文献中基于相容矩阵的不完备信息系统属性约简算法,提出了一种新的相容矩阵定义,该矩阵能够减少关注的条件属性。同时,也证明利用该矩阵所求的属性约简与相关文献中的属性约简是等价的。给出了其属性约简算法,提高了属性约简的速度。In rough set theory,knowledge reduction has been one of the key contents.In this paper a new tolerance matrix is defined relative to tolerance relation-based attribute reduction algorithm under incomplete information systems in related reference.It is able to reduce the number of conditional attributes which need to be concerned.At the same time, it is proved that the attribute reduction acquired from this new method is equivalent to the attribute reduction in related reference. An algorithm for attribute reduction is presented,the speed of attribute reduction in incomplete information systems is increased.

关 键 词：粗糙集 属性约简 不完备信息系统 相容矩阵 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]