基于区分向量的不完备决策系统属性约简算法  

Attribute Reduction for Incomplete Decision Systems Based on Discernibility Vector

作　　者：徐志海[1] 杨海峰[1] 

机构地区：[1]太原科技大学计算机科学与技术学院,太原030024

出　　处：《太原科技大学学报》2009年第1期12-16,共5页Journal of Taiyuan University of Science and Technology

基　　金：山西省自然科学基金(2006011041)

摘　　要：区分矩阵是属性约简的一种有效方法。针对不完备的决策系统,定义了区分矩阵的概念,给出了区分矩阵构造方法,并根据区分矩阵稀疏性的特点,定义了区分向量的概念,在此基础上,给出了一种基于区分向量的不完备决策系统属性约简算法。最后,采用恒星光谱数据,实验验证了该算法的有效性。Discernibility matrix based on rough set theory is an important content for attribute reduction of information system. In this paper, a new concept of discernibility matrix to incomplete decision systems is proposed, and the construction method of discernibility matrix is given. The concept of discernibility vector is proposed based on the, sparsity of discernibility matrix. An algorithm of attribute reduction of decison system is further proposed and its efficiency is verified through experiments.

关 键 词：粗糙集 区分矩阵 不完备决策系统 区分向量 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]