粗糙集近似与信息粒度

Rough Set Approximations and Information Granules

用粗糙集近似描述了三类常见信息系统(即Pawlak信息系统、不完备信息系统、不完备模糊信息系统)中对象的基本信息粒度。通过信息系统中对象属性值关于对象属性近似空间的上近似可以得到与对象具有相同或相似信息的对象集,即利用对象属性值关于对象属性近似空间的上近似将对象属性值信息变换成为对象的基本信息粒度。所得结论对信息系统中基本信息粒度的物理意义有了比较清楚和更加合理的解释。

In this paper, we prove that basic information granules in Pawlak information systems, incomplete information systems, and incomplete fuzzy information systems can be respectively described as three types of rough set approximations. By using the upper approximations of attribute value of an object with respect to the object-attribute approximation space, we can obtain an object set in which all the objects have the same or similar information, that is, we can transform the object-attribute value into basic intormation.granule tor each object. The obtained results give clear and reasonable explanation of the physical meaning of basic information granules.