可调节多粒度粗糙集的细化与粗化  

Refining or Coarsening Attribute Values in Adjustable Multigranulation Rough Set

作　　者：梁美社 谷怡欣 米据生[2] 靳晨霞 LIANG Meishe;GU Yixin;MI Jusheng;JIN Chenxia(Department of Mathematics and Physics,Shijiazhuang Tiedao University,Shijiazhuang 050043,China;School of Mathematics Science,Hebei Normal University,Shijiazhuang 050024,China;School of Economics and Management,Hebei University of Science&Technology,Shijiazhuang 050018,China)

机构地区：[1]石家庄铁道大学数理系,石家庄050043 [2]河北师范大学数学科学学院,石家庄050024 [3]河北科技大学经济管理学院,石家庄050018

出　　处：《人工智能科学与工程》2023年第9期1-11,共11页Journal of Southwest China Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(62076088,72101082),河北省自然科学基金项目(A2020208004),石家庄铁道大学研究生创新基金项目(YC2023246)。

摘　　要：在经典的多粒度粗糙集模型中,属性值细化和粗化后的下近似、上近似变化并不明显,因此在处理实际问题时存在一定局限性。该文首先构造了属性值细化和粗化时的可调节多粒度粗糙集,讨论了可调节多粒度粗糙集下、上近似的一些性质,并设计了相应的增量算法。最后在实验部分中,通过调节数据集、属性集的大小对增量算法和原有的静态算法的计算时间进行比较。实验结果表明提出的可调节多粒度粗糙集在细化和粗化时的增量算法可以更加准确、高效地处理大型数据。In the classical multigranulation rough set model,the change of upper and lower approximations after attribute values refining and coarsening are not obvious,so there are some limitations in dealing with practical problems.In this paper,the authors first construct the adjustable multigranulation rough set when the attribute values are refined and coarsened,discuss some properties of the upper and lower approximations of the adjustable multigranulation rough set,and design the corresponding incremental algorithm.Finally,in the experimental section,the calculation times of the incremental algorithm and the original static algorithm are compared by adjusting the sizes of the dataset and attribute set.The results demonstrate that the proposed incremental algorithm for adjustable multigranulation rough sets in refining and coarsening are more effective and accurate when handling large data.

关 键 词：多粒度 粗糙集 可调节多粒度粗糙集 增量算法 下近似 上近似 

分 类 号：O159[理学—数学]