A Novel Hierarchical Deep Matrix Completion Method
The matrix completion technique based on matrix factorization for recovering missing items is widely used in collaborative filtering, image restoration, and other applications. We proposed a new matrix completion model called hierarchical deep matrix completion (HDMC), where we assume that the varia...
Guardado en:
Autores principales: | , , , , , , , |
---|---|
Formato: | article |
Lenguaje: | EN |
Publicado: |
IEEE
2021
|
Materias: | |
Acceso en línea: | https://doaj.org/article/3f365e77f52f4a93b4a52aa3a2aac4e9 |
Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
id |
oai:doaj.org-article:3f365e77f52f4a93b4a52aa3a2aac4e9 |
---|---|
record_format |
dspace |
spelling |
oai:doaj.org-article:3f365e77f52f4a93b4a52aa3a2aac4e92021-11-19T00:04:37ZA Novel Hierarchical Deep Matrix Completion Method2169-353610.1109/ACCESS.2021.3049297https://doaj.org/article/3f365e77f52f4a93b4a52aa3a2aac4e92021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9313993/https://doaj.org/toc/2169-3536The matrix completion technique based on matrix factorization for recovering missing items is widely used in collaborative filtering, image restoration, and other applications. We proposed a new matrix completion model called hierarchical deep matrix completion (HDMC), where we assume that the variables lie in hierarchically organized groups. HDMC explicitly expresses either shallow or high-level hierarchical structures, such as taxonomy trees, by embedding a series of so-called structured sparsity penalties in a framework to encourage hierarchical relations between compact representations and reconstructed data. Moreover, HDMC considers the group-level sparsity of neurons in a neural network to obtain a pruning effect and compact architecture by enhancing the relevance of within-group neurons while neglecting the between-group neurons. Since the optimization of HDMC is a nonconvex problem, to avoid converting the framework of the HDMC models into separate optimized formulations, we unify a generic optimization by applying a smoothing proximal gradient strategy in dual space. HDMC is compared with state-of-the-art matrix completion methods on applications with simulated data, MRI image datasets, and gene expression datasets. The experimental results verify that HDMC achieves higher matrix completion accuracy.Yaru ChenXiaohong GuConghua ZhouXiaolong ZhuYi JiangJohn Kingsley ArthurEric Appiah ManteyErnest Domanaanmwi GanaaIEEEarticleMatrix completionhierarchical relationstructured sparsityregulationneural networkElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 7908-7920 (2021) |
institution |
DOAJ |
collection |
DOAJ |
language |
EN |
topic |
Matrix completion hierarchical relation structured sparsity regulation neural network Electrical engineering. Electronics. Nuclear engineering TK1-9971 |
spellingShingle |
Matrix completion hierarchical relation structured sparsity regulation neural network Electrical engineering. Electronics. Nuclear engineering TK1-9971 Yaru Chen Xiaohong Gu Conghua Zhou Xiaolong Zhu Yi Jiang John Kingsley Arthur Eric Appiah Mantey Ernest Domanaanmwi Ganaa A Novel Hierarchical Deep Matrix Completion Method |
description |
The matrix completion technique based on matrix factorization for recovering missing items is widely used in collaborative filtering, image restoration, and other applications. We proposed a new matrix completion model called hierarchical deep matrix completion (HDMC), where we assume that the variables lie in hierarchically organized groups. HDMC explicitly expresses either shallow or high-level hierarchical structures, such as taxonomy trees, by embedding a series of so-called structured sparsity penalties in a framework to encourage hierarchical relations between compact representations and reconstructed data. Moreover, HDMC considers the group-level sparsity of neurons in a neural network to obtain a pruning effect and compact architecture by enhancing the relevance of within-group neurons while neglecting the between-group neurons. Since the optimization of HDMC is a nonconvex problem, to avoid converting the framework of the HDMC models into separate optimized formulations, we unify a generic optimization by applying a smoothing proximal gradient strategy in dual space. HDMC is compared with state-of-the-art matrix completion methods on applications with simulated data, MRI image datasets, and gene expression datasets. The experimental results verify that HDMC achieves higher matrix completion accuracy. |
format |
article |
author |
Yaru Chen Xiaohong Gu Conghua Zhou Xiaolong Zhu Yi Jiang John Kingsley Arthur Eric Appiah Mantey Ernest Domanaanmwi Ganaa |
author_facet |
Yaru Chen Xiaohong Gu Conghua Zhou Xiaolong Zhu Yi Jiang John Kingsley Arthur Eric Appiah Mantey Ernest Domanaanmwi Ganaa |
author_sort |
Yaru Chen |
title |
A Novel Hierarchical Deep Matrix Completion Method |
title_short |
A Novel Hierarchical Deep Matrix Completion Method |
title_full |
A Novel Hierarchical Deep Matrix Completion Method |
title_fullStr |
A Novel Hierarchical Deep Matrix Completion Method |
title_full_unstemmed |
A Novel Hierarchical Deep Matrix Completion Method |
title_sort |
novel hierarchical deep matrix completion method |
publisher |
IEEE |
publishDate |
2021 |
url |
https://doaj.org/article/3f365e77f52f4a93b4a52aa3a2aac4e9 |
work_keys_str_mv |
AT yaruchen anovelhierarchicaldeepmatrixcompletionmethod AT xiaohonggu anovelhierarchicaldeepmatrixcompletionmethod AT conghuazhou anovelhierarchicaldeepmatrixcompletionmethod AT xiaolongzhu anovelhierarchicaldeepmatrixcompletionmethod AT yijiang anovelhierarchicaldeepmatrixcompletionmethod AT johnkingsleyarthur anovelhierarchicaldeepmatrixcompletionmethod AT ericappiahmantey anovelhierarchicaldeepmatrixcompletionmethod AT ernestdomanaanmwiganaa anovelhierarchicaldeepmatrixcompletionmethod AT yaruchen novelhierarchicaldeepmatrixcompletionmethod AT xiaohonggu novelhierarchicaldeepmatrixcompletionmethod AT conghuazhou novelhierarchicaldeepmatrixcompletionmethod AT xiaolongzhu novelhierarchicaldeepmatrixcompletionmethod AT yijiang novelhierarchicaldeepmatrixcompletionmethod AT johnkingsleyarthur novelhierarchicaldeepmatrixcompletionmethod AT ericappiahmantey novelhierarchicaldeepmatrixcompletionmethod AT ernestdomanaanmwiganaa novelhierarchicaldeepmatrixcompletionmethod |
_version_ |
1718420664154587136 |