粗糙集决策表边界属性分析

Analysis of boundary region attribute on rough set decision table

上近似、下近似是粗糙集理论的基本概念,由上下近似概念可导出粗糙集的正域、负域、边界等概念。对于决策信息系统,决策属性的等价类可以用粗糙集理论的上下近似来刻画,边界反映了其粗糙性。分析决策信息系统的边界类属性,结合可变精度粗糙集与经典粗糙集理论,对比边界类属性与正域类属性及负域类属性的差别,提出了正向迁移属性和负向迁移属性概念,结合例子给出正向迁移属性与负向迁移属性的求法,并对其意义加以说明。

The lower approximation and upper approximation are the basic concepts of rough set theory,associated with those important concepts such as positive region,negative region and boundary region.For the decision table,the equivalence classes determined by properties can also be characterized by lower approximation and upper approximation.And boundary region can be used to describe the rough nature of the equivalence classes as well.Combined with available precision rough set theory and classics rough set theory,this paper analyzes the boundary region attributes of decision table,and proposes the concepts of positive migration attribute and negative migration attribute by comparing positive region attributes and negative region attributes.An example is given to explain how to get the positive region attributes and negative region attributes.And the meaning of positive region attributes and negative region attributes is given simultaneously.