一种基于rough集的属性约简的改进算法  

An improved algorithm for attribute reduction based on rough sets

作　　者：李永华[1] 蒋芸[1] 王小菊[1] 

机构地区：[1]西北师范大学数学与信息科学学院,兰州730070

出　　处：《计算机应用》2008年第8期2000-2002,共3页journal of Computer Applications

基　　金：国家自然科学基金资助项目(60573096);西北师范大学2006-2010年度重点学科“网络计算”资助项目

摘　　要：目前粗糙集属性约简基本上是通过差别矩阵先求出属性核,然后在属性核的基础上再求出属性约简。这种计算方法具有较高的复杂度。因此提出并分析了属性的加权平均重要性的属性约简算法,该算法可以确保得到决策表的一个约简,且不需要计算核,减少计算量,提高计算速度。以实例验证了算法的正确性。At present, basically all attribute reduction based on Rough Sets is to extract the attribute core through the discernbility matrix, then calculate attribute reduction, but the method is still complex. This paper proposed and analyzed an attribute reduction algorithm of weighted mean attribute significance. The algorithm not only can get a reduction, but also does not need core calculating, It can reduce the computing effort and improve the computing efficiency. The algorithm is verified with the instance.

关 键 词：粗糙集 加权平均属性重要性 属性约简算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]