拟单层覆盖粗糙集中近似集的增量更新算法

拟单层覆盖粗糙集与集值信息系统之间存在一一对应的映射关系，当集值信息系统中的对象集动态添加或移除时，对应拟单层覆盖粗糙集中的信息单元也会随之改变，导致拟单层覆盖粗糙集中的近似集发生变化。针对拟单层覆盖粗糙集中近似集的动态更新问题，将拟单层覆盖粗糙集与增量学习相结合，提出近似集的增量更新算法。设计拟单层覆盖集中信息单元的更新算法，以分析信息单元的变化情况，分别构建近似集中可靠单元和争议单元的相关可靠单元集的更新算法。在此基础上，设计与可靠单元和争议单元更新算法相对应的增量更新算法，并且分析其时间复杂度。在UCI数据集上的实验结果表明，与静态算法相比，该算法在对象集发生添加和移除情况下的近似集更新效率分别提高21.5和29倍，能够有效提高近似集的计算效率。

Incremental Updating Algorithm for Approximation Sets on Semi-Monolayer Cover Rough Sets

A one-to-one mapping relationship exists between the semi-monolayer cover rough sets and the set-valued information system.When an object set in a set-valued information system is dynamically added or removed, the information unit that corresponds to the semi-monolayer cover rough sets also changes.This leads to the change in the approximation sets on the semi-monolayer cover rough sets.To address the dynamic updating problem of approximation sets on semi-monolayer cover rough sets, an incremental updating algorithm for approximation sets is proposed by combining semi-monolayer cover rough sets with incremental learning.The updating algorithom, comprising a centralized information unit with semi-monolayer coverage, is designed to analyze the changes in the information unit.The updating algorithms of relevant reliable unit sets of approximation centralized reliable and controversial units are constructed.An incremental updating algorithm that corresponds to the updating algorithms of reliable and controversial units is designed, and then its time complexity is analyzed.The experimental results on the UCI dataset indicate that compared with the static algorithm, the update efficiency of the approximation sets with the addition and removal of the object sets is improved by 21.5 and 29 times, respectively.Therefore, the incremental updating algorithm can effectively improve the computational efficiency of the approximation sets.