一般关系粗糙集的近似分类精度和质量  

APPROXIMATION CLASSIFIED PRECISION AND QUALITY BASED ON GENERAL RELATION ROUGH SET

作　　者：朱晓钟[1,2] 杨勇[1] 朱英丽[1] 

机构地区：[1]西北师范大学数学与信息科学学院,甘肃兰州730070 [2]河海大学计算机与信息学院(常州),江苏常州213011

出　　处：《计算机应用与软件》2011年第5期52-54,共3页Computer Applications and Software

基　　金：国家自然科学基金(10771171);兰州市科技计划项目(2008-1-34)

摘　　要：由等价关系R所决定的近似空间(U,R)上,近似分类精度指出了可能的决策中正确决策所占的百分比,近似分类质量则表明了应用知识R能确切地划入已知分类的对象的百分比。将近似分类精度和分类质量的概念推广到一般(二元)关系所决定的广义近似空间上。通过引入独立集概念,给出了度量公式,最后通过实例验证了其合理性。In approximation space（U,R） determined by equivalence relation R,the approximate classification accuracy expresses the percentage the correct decision occupies among all possible decisions made,and the approximate classification quality indicates the percentage the applied knowledge R precisely belongs to the objects of known classification.In this article these two mentioned concepts are extended to the generalised approximation space which is determined by general（binary） relation.By introducing the concept of solitary set,the metric formulae are given,and their rationalities are attested with an example at the end.

关 键 词：粗糙集 一般(二元)关系 近似分类精度 近似分类质量 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]