基于二进制可辩识矩阵变换的离散化算法  

Discretization algorithm in rough set based on binary discernibility matrix's tansformation

作　　者：侯利娟[1] 颜宏文[1] 

机构地区：[1]长沙理工大学计算机与通信工程学院,湖南长沙410076

出　　处：《计算机工程与设计》2008年第9期2330-2332,共3页Computer Engineering and Design

基　　金：湖南省教育厅基金项目(06C125)

摘　　要：在用粗糙集理论处理决策表进行约简时,要求决策表中的各值用离散值表达,即离散化。求最小数目的断点集是一个NP-hard问题,解决这类问题的一般方法是采用启发式算法求出最优或次优解,给出了离散化中的二进制可辩识矩阵的定义,并提出了基于二进制可辩识矩阵变换的离散化算法,实例证明,该算法是有效的和高效的。Before processing decision table＇s reduction in rough set, it request the data is discretized. In other words, it requests discretization. The NP-hard problem is proved for seaching the minimum cut set. The common method to solve this problem is to adopt the heuristic algorithm. The definition of binary discernibility in discretization fields is given and the algorithm of discretization based on binary discernibility matrix＇s tansformation is proposed, the example proved this algorithm is effective and highly efficient .

关 键 词：粗糙集理论 决策表 二进制可辨识矩阵 离散化 断点核 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]