论文部分内容阅读
针对粗糙集理论中计算决策表核属性时出现的若干问题,讨论并分析了这些问题产生的原因.在此基础上,给出了不同目标下的广义约简定义,提出了必要区分关系和非必要区分关系等概念,由此给出了一般意义下的分辨矩阵定义.根据广义约简的定义,将已有的改进差别矩阵求核方法与早期Skowron等人提出的经典分辨矩阵求核方法作了对比,证明了两种方法完全相同.
Aiming at some problems that arise when calculating the nuclear attribute of decision table in rough set theory, the causes of these problems are discussed and analyzed.On this basis, the definition of generalized reductions under different objectives is given, and the necessary distinction between the two is proposed The definition of the discernibility matrix in general sense is given.According to the definition of generalized reduct, the existing improved differential matrix method and the method of kernel discerning of classical discernibility matrix proposed by Skowron et al. In contrast, it proves that the two methods are exactly the same.