摘要
为了拓展多粒度粗糙集理论在覆盖近似空间上的研究,本文利用元素的最小描述并集并结合条件概率,提出了3种多粒度覆盖粗糙集模型。在模型定义基础上,本文研究了3种新模型的一些特有性质,探讨了新模型与一些已有模型的内在联系与区别,对3种新模型进行了比较。研究结果表明一些已有模型是本文模型的特殊形式,是已有模型的有效拓展。
To generalize m u ltig ra n u la tio n rough sets to a covering-based a p pro xim atio n sp a ce ,th is paper proposesthree k in d s of covering-based m u ltig ra n u la tio n rough sets by em p lo ying the c o n d itio n a l p ro b a b ility between the targetconcept and the u n io n of the m in im a l descrip tions of elem ents. Based on new d e fin itio n s ,some basic properties ofthese m odels w ere investigated and th e ir relatio nships w ith some e xistin g covering-based m u ltig ra n u la tio n rough setsare reve aled. The in te r-re la tio n s h ip am ong the three new m odels is also e x p lo re d . The discussions show that the p ro posedm odels are a special form of text model, as well as extensions of some e xistin g covering-based multigranulation rough sets.
出处
《智能系统学报》
CSCD
北大核心
2016年第4期534-538,共5页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金项目(61305052
61403329
61663002)
关键词
粗糙集
多粒度
条件概率
覆盖
最小描述
rough sets
multigranulation
conditionalprobability
covering
minimal description
