论文部分内容阅读
以同时具有遗漏型和缺席型未知属性值的不完备信息系统为研究对象,分析了由特征关系所得到的特征类中并非任意两个元素都满足容差关系这一不足,将最大相容块技术引入到特征类中,提出了特征相容块的概念,不仅对基于特征相容块的粗糙集的基本性质进行了讨论,而且将特征相容块粗糙集与特征关系粗糙集进行了对比分析,研究结果表明,相比于特征关系,采用特征相容块的方法,可以得到更大的下近似和更小的上近似。
Taking the incomplete information system which has both missing and absent-type unknown attribute values as the research object, it is analyzed that the feature class obtained from the feature relationship does not satisfy the tolerance relation that any two elements satisfy the tolerance relation, and the largest compatible block This paper introduces the concept of feature-compatible block, not only discusses the basic properties of rough set based on feature-compatible block, but also compares the feature-compatible block rough set with the feature-relation rough set The results show that, compared with the feature relation, the method of feature-compatible block can get larger lower approximation and smaller upper approximation.