不完备信息系统中的可变精度分类粗糙集模型  

Rough set model based on variable parameter classification in incomplete information systems

作　　者：杨习贝[1] 杨静宇[1] 於东军[1] 吴陈[2] 

机构地区：[1]南京理工大学计算机科学与技术学院,南京210094 [2]德瑞索大学信息科学与技术学院,美国费城PA19104

出　　处：《系统工程理论与实践》2008年第5期116-121,共6页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(60472060,60572034,60632050);江苏省自然科学基金(BK2006081)

摘　　要：在不完备信息系统中,容差关系过于宽松,而相似关系则过于严格.根据这样的解释,提出了一种新的基于可变精度分类的拓展粗糙集模型,其中的分类方式相比较于容差关系和相似关系显得更为灵活,是这两者的一种广义化表现形式,且可变精度分类也是限制容差关系的一种改进形式.在此基础上,将这种拓展粗集模型与基于容差关系和相似关系的拓展粗集模型进行了对比分析.最后在不完备信息系统中使用新的拓展粗集模型讨论了确定和可能性规则的直接生成方法,并进行了实例分析以说明其有效性.In incomplete information systems, the tolerance relation is too loose while the similarity relation is too strict. Aeeording to these explanations, a new kind of rough set model that is based on the variable parameter classification of objects is proposed. The variable parameter classification is a kind of generalized classification form and it is more flexible than tolerance and similarity relations in practical applications. Furthermore, it can be regarded as an improvement of the limited tolerance relation. From what have been discussed above, the new-defined generalized rough set model is compared with expanded rough sets those are on the basis of tolerance and similarity relations. Finally, the immediate method of certain and possible rules＇ induction is studied by the new expanded rough set model, an illustrative example is used to show its validity.

关 键 词：不完备信息系统 容差关系 相似关系 可变精度分类 粗糙集 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]