摘要： 为解决粗糙集中的属性约简问题，提出一种完备的最小属性约简方法。将差别矩阵中所有有关属性区分的信息都浓缩进一个差别向量组，计算每个属性在区分2个对象的属性集合中出现的概率，作为属性重要性的启发式信息，建立最小属性约简树，得到属性约简。分析结果表明，该方法可以获得所有的最小属性约简。

关键词: 粗糙集, 决策表, 差别属性集, 差别向量组, 最小属性约简树, 最小属性约简

Abstract: Attribute reduction is the basic problem of rough sets theory. A method for minimal attributes reduction in consistent decision table is proposed in this paper. The discernible information in consistent decision tables is described with discernible vector array. A minimal attribute reduction tree is generated based on the probability of the attributes which discern two objects. All minimal attribute reductions are got from minimal attributes reduction tree. The result of the method is proved to be complete and minimal.

Key words: rough set, decision table, discernible attribute set, discernible vector array, minimal attribute reduction tree, minimal attribute reduction

中图分类号: 

• TP311.12