摘要
基于变精度粗糙集理论与包含度理论,引入了不协调目标信息系统的上、下分布约简的概念,并讨论了它们之间的关系.上(下)分布约简是保持每个决策类的上(下)近似不变的最小属性集,由约简系统产生的命题规则与由原系统产生的命题规则是相容的,即约简不会改变由对象所产生的规则的决策结果.通过对这两种知识约简的等价刻画,得到了上、下分布知识约简的判定定理和可辨识属性矩阵,从而提供了不协调目标信息系统知识约简的新方法.
The main objective of the paper is to introduce some new concepts of knowledge reduction based on variable precision rough set theory such as upper distribution reduction and lower distribution reduction. The decision rules derived from the distribution consistent set are compatible with the ones derived from original system. Their equivalent definitions are studied. The relationships among alternative reducts in inconsistent information systems are discussed. The judgement theorems and discernibility matrixes with respect to upper and lower reductions are obtained. So one can calculates the reducts by the discernibility formulas. These results are meaningful both in the theory and in applications.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2004年第1期76-82,共7页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(10271039)
973项目(2002CB312206)
关键词
变精度粗糙集
知识约简
信息系统
协调集
variable precision rough set
information system
knowledge reduction
consistent set
