摘要
本文分析粗糙集中不同正区域之间的定量关系,给出正区域和近似精度计算的一个简洁的递归公式.在一类扩展型属性约简算法中(如Hu算法,Jelonek算法等),应用该递归公式来完成大量的正区域或近似精度以及与之相关的属性重要性的计算,可以较大幅度地减少计算量,提高属性约简算法的速度.
In this paper, the relationship between positive regions in the context of rough set is anayzed, and then on the basis of which a simple recursive formula for computing the positive region and the approximation quality is derived. It turns out that for a class of expansion-type attribute reduction algorithms, the aplication of the recursive formula can efficiently reduce the computational effort on the frequent computation of the positive region or approximation qualities and of significant values of attributes, thus increasing the speed of reduct-finding process.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2002年第4期474-477,共4页
Pattern Recognition and Artificial Intelligence
基金
教育部科学技术研究重点项目(No.00185)
福建省自然科学基金(No.A0010009)