Graph regularized nonnegative matrix factorization with label discrimination for data clustering

Non-negative Matrix Factorization (NMF) is an effective method in multivariate data analysis, such as feature learning, computer vision and pattern recognition. For practical clustering tasks, NMF ignores both the local geometry of data and the discriminative information of different classes. In thi...

Full description

Saved in:
Bibliographic Details
Published in:Neurocomputing (Amsterdam) 2021-06, Vol.440, p.297-309
Main Authors: Xing, Zhiwei, Ma, Yingcang, Yang, Xiaofei, Nie, Feiping
Format: Article
Language:English
Subjects:
Clustering
Laplacian matrix
Non-negative matrix factorization
Semi-supervised learning
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Non-negative Matrix Factorization (NMF) is an effective method in multivariate data analysis, such as feature learning, computer vision and pattern recognition. For practical clustering tasks, NMF ignores both the local geometry of data and the discriminative information of different classes. In this paper, we propose a new NMF method under graph and label constraints, named Graph Regularized Nonnegative Matrix Factorization with Label Discrimination (GNMFLD), which attempts to find a compact representation of the data so that further learning tasks can be facilitated. The proposed GNMFLD jointly incorporates a graph regularizer and the prior label information as additional constraints, such that it can effectively enhance the discrimination and the exclusivity of clustering, and improve the clustering performance. Empirical experiments demonstrate the effectiveness of our new algorithm through a set of evaluations based on real-world applications.
ISSN:0925-2312
1872-8286
DOI:10.1016/j.neucom.2021.01.064