基于模糊相容关系的集值决策表的相对约简

从模糊粗糙集的角度讨论集值决策表的相对约简。首先,基于2个对象取相同值的可能性大小,在集值信息系统中定义了一个模糊相容关系,给出了模糊相容关系的上近似和下近似;其次,通过引入模糊近似质量,定义了集值决策表中条件属性的重要度与相对重要度,进而提出了一种计算相对约简的启发式算法;最后,通过实例说明该算法能够得到集值决策表的相对约简。

Relative Attribute Reduction of Set-valued Decision Table Based on Fuzzy Tolerance Relation

The purpose of this paper is to study relative attribute reduction of set-valued decision table from the viewpoint of fuzzy rough theory. Firstly, the authors define a fuzzy tolerance relation based on the possibility of two objects having the same value in such information systems with respect to the condition attributes and give the upper approximation and lower approximation of fuzzy tolerance relation. Secondly, by introducing fuzzy approximation quality of the set-valued decision table, the paper defines the significance and relative significance of condition attributes, and proposes a heuristic algorithm for computing relative reduction. Finally, the example shows that this algorithm can obtain the relative reduction of a set-valued decision table.