摘要
目前有关粒计算的理论与方法主要有商空间理论、词计算理论和粗糙集理论。以粗糙集理论发展的粒计算理论为基础,定义了幂图、粒度幂图及二进制粒等概念,提出基于二进制粒计算与粒度幂图的两种属性约简算法,把属性约简问题转化为在粒度幂图中的搜索问题,为属性约简开辟了新的途径。理论分析表明,所提出的算法是有效可行的。
At present ,there are three theories about granular computing,Theory of Quotient Space ,Theory of Computing with Words and Theory of Rough Set. Based on Granular Computing Theory of Rough Set, power graph ,granularity-power graph and binary granules are defined. Two attribute reduction algorithms based on binary granular computing and granularity-power graph are put forward ,which translate question of attribute reduction into problem of searching in granularity-power graph. These algorithms provide a new method in attribute reduction. Theoretical analysis shows that these algorithms of this paper are efficient and feasible.
出处
《广西师范大学学报(自然科学版)》
CAS
北大核心
2008年第2期81-84,共4页
Journal of Guangxi Normal University:Natural Science Edition
基金
国家自然科学基金资助项目(60475019,60775036)
2006年博士学科点专项科研基金资助项目(20060247039)
关键词
粗糙集
粒计算
属性约简
幂图
二进制粒
rough set
granular computing
attribute reduction
power graph
binary granules