摘要
属性约简是粗糙集理论进行数据挖掘的基本途径,相关算法主要基于核。而差别矩阵在求核计算中具有重要意义,且现实世界变化很快,信息系统中的对象在不断动态变化,已得到的核将可能不再有效,这就需要对已得到的核进行更新。本文讨论了信息系统中新增对象的各种情况,提出了一个基于分块差别矩阵的增量式求核算法,并通过实例和实验验证了该算法的有效性。
Attribute reduction is the basic method of data mining in rough set theory,and the correlation algorithm is mainly based on core.The discernibility matrix has important significance in calculating the core.The real world changes rapidly,and the objects in the information system are constantly changing dynamically.The core received may no longer be valid,so we need to dynamically modify the core.This article discusses the analysis of new objects in information systems,and an incremental updating algorithm based on block discernibility matrix is proposed.And the validity of the algorithm is verified by examples and experiments.
作者
左芝翠
莫智文
ZUO Zhi-cui;MO Zhi-wen(College of General Education,Southwest University of Science and Technology City College,Mianyang 621000;College of Mathematics and Software Science,Sichuan Normal University,Chengdu 610068.China;Institute of Intelligent Information and Quantum Information,Sichuan Normal University,Chengdu 610068,China)
出处
《模糊系统与数学》
北大核心
2022年第5期166-174,共9页
Fuzzy Systems and Mathematics
基金
国家自然科学基金资助项目(11671284)
四川省科技计划项目(2017JY0197)。
关键词
粗糙集
属性约简
核
分块差别矩阵
增量式更新
Rough Set
Attribute Reduction
Core
Block Discernibility Matrix
Incremental Update
