一种实值属性信息系统的粗集约简方法

Rough Set Rules Reduction for Real Value Information Systems

本文研究应用粗集理论对实值信息系统属性进行约简的方法 .对实值属性信息系统进行约简的根本问题是如何对实值属性离散化 .通过对离散化方法与属性约简的关系进行研究 ,提出了实值属性离散化的一种自动确定属性类别的方法 ,并结合粗集理论给出了对实值属性信息系统约简的算法 .用所提出的算法进行了实验 。

In this paper, we study the problem of rules reduction for real value information systems by using of rough set theory. Existing methods for rules reduction are applicable to unreal value knowledge systems, not adequate to highly effective large scale data processing especially for real value knowledge systems. Based on the theory of rough set, a method for rules reduction to real value knowledge systems is proposed in this paper. Experimental results show the method is feasible and effective both for real value knowledge systems and unreal value knowledge systems. The method gives a way to effectively deal with knowledge systems for rules reduction.