摘要
经典粗集理论以等价关系为基础,它们在处理不完备信息系统的时候存在一定的局限性.作者首先从扩展粗集理论适应范围的角度出发,引入信息系统对象的先验概率,结合量化相容关系的思想,提出基于先验概率的相容关系模型,然后以该模型为基础研究了不完备信息系统的属性约简,通过严格的数学证明,得到基于分辨矩阵的属性约简算法.最后通过一个实例验证了该改进相容关系模型及其基于分辨矩阵的约简算法.表1,参20.
The classic rough set theory is based on equivalence relation, but it is not able to handle the incomplete information system. First, an extended rough set model based on valued tolerance relation and prior probability obtained out of incomplete information system is founded. Then, as a part of the model, the attribute reduction of incomplete information system and the corresponding discernibility matrix are proposed, and the attribute reduction is obtained by the normal form transform. Finally, the model and the distribution reduction method are verified under an incomplete information system. 1 tab., 20refs.
出处
《湖南科技大学学报（自然科学版）》
CAS
北大核心
2009年第3期73-77,共5页
Journal of Hunan University of Science And Technology：Natural Science Edition
基金
四川省科技计划项目(2008GZ0003)
四川省科技厅科技攻关项目(07GG006-014)
关键词
粗集
不完备信息系统
先验概率
属性约简
rough sets
incomplete information systems
prior probability
attribute reduction
