摘要
基于谱聚类的无监督特征选择主要涉及相关系数矩阵和聚类指示矩阵,在以往的研究中,学者们主要关注于相关系数矩阵,并为此设计了一系列约束和改进,但仅关注相关系数矩阵并不能充分学习到数据内在结构.考虑群组效应,本文向聚类指示矩阵施加F范数,并结合谱聚类以使相关系数矩阵学习更为准确的聚类指示信息,通过交替迭代法求解两个矩阵.不同类型的真实数据集实验表明文中方法的有效性,此外,实验表明F范数还可以使方法更加鲁棒.
Unsupervised feature selection based on spectral clustering mainly involves the correlation coefficient matrix and the clustering indicator matrix.In previous studies,scholars have mainly focused on the correlation coefficient matrix,designing a series of constraints and improvements for it.However,focusing solely on the correlation coefficient matrix cannot fully learn the intrinsic structure of data.Considering the group effect,this study imposes the F-norm on the clustering indicator matrix and combines it with spectral clustering to make the correlation coefficient matrix learn more accurate clustering indicator information.The two matrices are solved through an alternating iteration method.Experiments on different types of real datasets show the effectiveness of the proposed method.In addition,experiments show that the F-norm can also make the method more robust.
作者
林清水
田鹏飞
张旺
LIN Qing-Shui;TIAN Peng-Fei;ZHANG Wang(Software College,Liaoning Technical University,Huludao 125105,China;Department of Basic Teaching,Liaoning Technical University,Huludao 125105,China)
出处
《计算机系统应用》
2024年第7期201-212,共12页
Computer Systems & Applications
关键词
无监督特征选择
谱聚类
群组效应
F范数
降维
unsupervised feature selection
spectral clustering
group effect
F-norm
dimension reduction
