一种改进的差别矩阵及其求核方法  

An Improved Discernibility Matrix and Approach to Computing a Core

作　　者：汪小燕[1] 

机构地区：[1]安徽工业大学计算机学院,安徽马鞍山243002

出　　处：《安徽工业大学学报（自然科学版）》2009年第1期86-88,95,共4页Journal of Anhui University of Technology（Natural Science）

摘　　要：属性约简是粗糙集理论中的重要研究内容之一,求核运算是属性约简的基础,直接影响着知识约简的效率。针对目前求核方法存在的问题,提出一种基于分布函数的改进的差别矩阵及属性核计算方法,并证明了方法的正确性,改进的差别矩阵不仅规模小,而且适用于任何决策表求核。Attribution reduction is one of the important topics in the research on rough set theory. The core obtaining lays a foundation in the attribute reduction, which affects the efficiency of attribute reduction. In order to correct some problems of computing the core of a decision table, an improved discernibility matrix definition based on distribution function together with a method for computing the core is put forward in this paper,and the correctness of this method is proved.The improved discernibility matrix is not only small in scale but also suitable for any decision tables for computing the core.

关 键 词：粗糙集 差别矩阵 核 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]