摘要
对传统的粗集约简方法进行了推广,讨论了在特定相容关系下的信息系统以及协调目标信息系统上辨识矩阵的构造方法,推广之后不再需要对论域形成划分,只需要形成覆盖.针对可辨识矩阵的概念推广到覆盖的情形,提出了3种新的基于覆盖广义粗集的约简并对其优良性质进行了讨论,进而在此基础上给出了近似覆盖约简的启发式算法.
In this paper, the traditional reduction methods in rough sets theory is extender to the situation in which covering rather than partition of all the objects are constructed. The construction method of discrimination under the given compatible relation and harmonic objective information system is discussed. The definition of discrimination matrixes is generalize to the situation of covering and proposed three new definitions on knowledge reduction based on covering generalized rough sets. The favorable properties of these reductions are discussed and heuristic algorithms of approximate covering are presented at the end of the paper base on these properties.
出处
《纺织高校基础科学学报》
CAS
2007年第2期198-202,共5页
Basic Sciences Journal of Textile Universities
关键词
信息系统
粗糙集
知识约简
近似覆盖约简
information system
rough sets
knowledge reduction
covering approximation reduction
