基于邻域粗糙集的在线流特征选择  

Online Streaming Feature Selection based on Neighborhood Rough Sets

作　　者：蔡晶晶 荀亚玲[1] 王林青 贺慧爱 孙晶晶 CAI Jing-jing;XUN Ya-ling;WANG Lin-qing;HE Hui-ai;SUN Jing-jing(School of Computer Science and Technology,Taiyuan University of Science and Technology,Taiyuan 030024,China)

机构地区：[1]太原科技大学计算机科学与技术学院,太原030024

出　　处：《太原科技大学学报》2025年第1期1-6,共6页Journal of Taiyuan University of Science and Technology

基　　金：国家自然科学基金(62272336);山西省研究生教育创新项目(2022Y699Z)。

摘　　要：针对现有基于粗糙集的在线流特征选择算法侧重于处理不相关和冗余特征,忽略了特征与特征之间的相关性,以及缺乏有效的动态更新机制,提出一种基于相关性邻域粗糙集的在线流特征选择算法OFSGN.使用欧氏距离计算特征之间的相似性,通过一种新定义的邻域粗糙集方法获取邻域集合,结合了特征与特征和标签之间的相关性,通过在线重要度分析和在线冗余度分析获取了最优特征子集。大量实验结果表明了该特征选择算法的有效性,且该算法在准确率方面具有明显的优势。For the existing online stream feature selection algorithms based on rough sets,they focus on the processing of irrelevant and redundant features,ignore the correlation between features and lack of effective dynamic updating mechanism.An online flow feature selection algorithm OFSGN based on correlation neighborhood rough set is proposed.Firstly,the algorithm calculates the similarity between features by using Euclidean distance.Then,a new neighborhood rough set method is used to obtain neighborhood sets,and the correlation between feature and feature and label is combined.Finally,the optimal feature subset is obtained by online significance analysis and online redundancy analysis.A large number of experimental results show the effectiveness of the feature selection algorithm,and the algorithm has obvious advantages in accuracy.

关 键 词：粗糙集 在线流 特征选择 邻域粗糙集 相关性 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]