目前,人们对广义粗糙集的研究主要集中在集合的近似计算上,而真正利用广义粗糙集进行属性约简的研究还很少见。基于广义粗糙集理论研究了信息系统中属性约简的问题。首先定义了关系信息系统及其属性约简的概念。然后给出了属性约简的判定定理和辨识矩阵并详细地研究了属性约简的基本性质。最后给出了计算信息系统中属性的所有约简的一般方法。
The existing researches on generalized rough sets are mainly concentrated on set approximations; less effort has been made for attribute reduction. Attribute reduction in information systems based on generalized rough sets is studied. Firstly, a relation information system and its attribute reduction are defined. A judgment theorem and a discernibility matrix associated with the attribute reduction are presented. Finally, a method to compute all the reductions in information systems is proposed.