一种基于区分度矩阵的属性约简及规则提取算法  

Attribute Reduction and Rule Extraction Algorithms Based on Discernibility Degree Matrix

作　　者：王杨[1] 

机构地区：[1]辽宁石油化工大学计算机与通信工程学院,抚顺113001

出　　处：《计算机与数字工程》2011年第9期7-9,48,共4页Computer & Digital Engineering

摘　　要：在Rough Set理论中属性最小约简及规则提取是NP-hard的。通过研究Rough Set理论中属性约简和值约简问题,提出了区分度矩阵的概念,同时利用矩阵中属性区分度的信息,提出了一种获取属性约简及规则提取的简化算法。实验结果表明,该算法是正确、有效、可行的。Attribute reduction and rule extraction are NP-hard for the rough set theory. In this paper, two important issues which are attribute reduction and value reduction were discussed. The definition of discernibility degree matrix was presented, and the novel algorithms for attribute reduction and rule extraction were given accordingly. Experimental results show the algorithms are effective and feasible.

关 键 词：ROUGH SET理论 属性约简 值约简 区分度矩阵 规则提取 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]