【摘要】 针对不确定信息系统属性约简问题,提出一种基于颗粒细度的属性约简方法。给出一个以粒细度为属性重要性的概念,并用此重要性作为启发信息进行属性的约简。从原始的不确定信息系统出发,定义一个利用粗糙集中的等价关系构建相容的粒度空间,从而缩减原始系统所占用的空间,避免计算时对象的重复。实现计算量的降低。更多还原

【Abstract】 This paper defines a concept of attribute significance based on the granular fineness,and uses the significance as the inspiration for attribute reduction.From the original uncertainty information systems,a space of granular based on equivalence relation of rough sets is defined.This method greatly reduces the space occupied by the original system to avoid duplication of the calculation object.And the computation scale can be decreased.更多还原