摘要
为了对动态变化的决策表进行属性约简处理,在改进的分辨矩阵的基础上,提出一种增量式属性约简算法,当决策表添加新的记录后,能快速得到新决策表的所有约简和最小约简.此外,通过对不相容决策表的正区域的决策值和边界域对原决策表进行分解,得到了一种分布式增量属性约简模型.仿真研究表明了算法的正确性和高效性.
Incremental algorithms for attribute reduction based on modified discernibility matrix are proposed, by which minimal attribute reduction cluster of new decision table can be obtained quickly when new records are added to primary decision table. A distributed model of incremental attribute reduction is also presented by decomposing values of decision attribute of positive region and boundary region in non-tolerant decision table. The simulation experiments show the validity and effectiveness of algorithms.
出处
《控制与决策》
EI
CSCD
北大核心
2007年第3期268-272,277,共6页
Control and Decision
基金
国家自然科学基金项目(60373111
60573068)
新世纪优秀人才支持计划项目(NCET)
关键词
粗集
属性约筒
增量式
分布式
Rough set
Attribute reduction
Incremental
Distributed
