%0 Journal Article
%T 基于核最小二乘回归子空间分割的高维小样本数据聚类<br>High dimension small sample data clustering using kernel least square regression subspace segmentation
%A 简彩仁
%A 翁
%A 谦
%A 陈晓云
%J 福州大学学报(自然科学版)
%D 2018
%R 10.7631/issn.1000-2243.16439
%X 针对基于线性表示理论的子空间分割方法没有考虑高维小样本数据的非线性性质，借鉴核理论，提出核最小二乘回归子空间分割方法，使子空间分割方法适合高维小样本数据的非线性性质. 经6个基因表达数据集和4个图像数据集上的实验，表明该方法是有效的.<br>The classical subspace segmentation methods based on linear representation theory do not consider the nonlinear properties of high dimension small sample data. In sight of the kernel theory，the kernel least square regression subspace segmentation method is proposed to make the subspace segmentation method suitable for the nonlinear properties of high dimension small sample data. Experiments on six gene expression datasets and four image datasets show that the method is effective
%K 最小二乘回归 子空间分割 核理论 聚类 高维小样本&lt
%K br&gt
%K least square regression subspace segmentation kernel theory clustering high dimension small sample
%U http://xbzrb.fzu.edu.cn/ch/reader/view_abstract.aspx?file_no=201801007&flag=1