云南省人工智能重点实验室

科研成果

当前位置：首页>科研成果>论文列表

论文作者：	Zhenqiu Shu ,Furong Zuo,Wenli Wu,Congzhe You.
论文名称：	Dual local learning regularized NMF with sparse and orthogonal constraints
论文发表刊物：	APPLIED INTELLIGENCE
论文发表时间：	2022
卷号页码：
论文描述：
收录情况：	SCI Indexed
论文摘要：
Non-negative matrix factorization (NMF) has shown remarkable competitiveness in the past few years. To fully exploit various known prior knowledge hidden in data, this paper proposes a dual local learning regularized NMF with sparse and orthogonal constraints (DLLNMF-SO) algorithm. DLLNMF-SO constructs two local learning regularizers to consider the geometric structure and discriminative information embedded in data and feature space, respectively. Besides, it makes full use of sparse selfrepresentation information by adding the l2,1-norm constraint. Meanwhile, the orthogonal constraint is imposed on the basis vectors to preserve the correspondence between samples and basic vectors. We give an efficient iterative updating scheme for the optimization problem of DLLNMF-SO and provides its convergence guarantee. We demonstrate that our proposed approach outperforms other competitors by conducting serval experiments on three benchmark datasets.
点击此处下载文章内容