论文部分内容阅读
Skowron分辨矩阵是代数观点属性约简模型的一种演化,其本质在于保持系统中非冲突对象与其他对象的可分辨关系不变,不能刻画常见的非代数观点属性约简准则.属性约简准则的本质体现为保持决策信息系统的某种特定可分辨特性不发生变化,决策信息系统具有多方面可分辨特性,单一属性约简准则仅能刻画其中某一特性.为将不同的属性约简准则运用统一的分辨矩阵形式加以描述,在定义条件等价类的决策向量基础上,构建了决策向量简化决策系统,进而设计满足不同属性约简准则的分辨矩阵及分辨函数,给出其与对应准则属性约简模型的等价性证明,推理证明与仿真实例说明了该方法的可行性与有效性.
Skowron discernibility matrix is an evolution of algebraic viewpoint attribute reduction model. The essence of this dissertation is that the discernible relationship between non-conflicting objects and other objects in the system remains unchanged, and the common non-algebraic attribute reduction rules can not be described. The essence of decision-making information system to maintain a particular discernable nature does not change, decision-making information system has a multi-faceted distinguishing feature, a single attribute reduction guidelines can only characterize one of them .In order to different attribute reduction guidelines Based on the decision vector of the condition equivalence classes, a decision-making vector simplified decision-making system is constructed, and then the discernibility matrix and resolution function that satisfy the different attribute reduction criteria are designed. The corresponding criteria The proof of equivalence of the attribute reduction model, the inference and simulation examples show the feasibility and effectiveness of the proposed method.