研究决策表中属性的重排。对同一个决策表,把其中不同的属性分别指定为结论属性,借助决策表中属性之间的约简和冗余关系判别哪种重排可以导致粗糙集以及哪种属性重排导致非粗糙集。提出了属性重排粗糙度的概念,基于属性重排粗糙度提出了属性最佳逻辑流向、最佳逻辑属性等概念。分析并指出不同属性重排所对应的粗糙度和属性最佳逻辑流向反映了决策表中不同属性之间的因果蕴含关系。数值算例表明,利用决策表的属性重排可以有效地对缺失数据进行插补。 Study the rearrangement of attributes in decision tables. For the same decision table, the different attributes are respectively designated as the conclusion attributes, and the reduction and redundancy relationship between the attributes in the decision table determine which kind of rearrangement can lead to the rough set and which attribute rearrangement leads to the non-rough set . The concept of attribute rearrangement roughness is proposed. Based on the attribute rearrangement roughness, the concepts of optimal attribute logic flow and best logical attribute are proposed. It is analyzed and pointed out that the optimal logic flow of the roughness and attribute corresponding to the rearrangement of different attributes reflects the causal implication of different attributes in the decision table. Numerical examples show that the use of decision table attribute rearrangement can effectively interpolate missing data.