A Novel Hierarchical Deep Matrix Completion Method

IEEE Access

卷:9

Xiaohong Gu¹  John Kingsley Arthur¹  Xiaolong Zhu²  Ernest Domanaanmwi Ganaa²  Yaru Chen²  Yi Jiang²  Conghua Zhou³  Eric Appiah Mantey³ 

[1] Pneumology Department, Wuxi Children&x2019;

[2] School of Computer Science and Communication Engineering, Jiangsu University, Jiangsu, China;

[3] s Hospital, Jiangsu, China;

关键词: Matrix completion;    hierarchical relation;    structured sparsity;    regulation;    neural network;   

DOI  :  10.1109/ACCESS.2021.3049297

来源: DOAJ

【 摘 要 】

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.

【 授权许可】

Unknown