多核低冗余表示学习的稳健多视图子空间聚类方法

doi:10.11959/j.issn.1000-436x.2021217

通信学报 ›› 2021, Vol. 42 ›› Issue (11): 193-204.doi: 10.11959/j.issn.1000-436x.2021217

多核低冗余表示学习的稳健多视图子空间聚类方法

李骜¹^,², 王卓¹, 于晓洋², 陈德运¹, 张英涛³, 孙广路¹

¹ 哈尔滨理工大学计算机科学与技术学院，黑龙江哈尔滨 150080
² 哈尔滨理工大学仪器科学与技术博士后流动站，黑龙江哈尔滨 150080
³ 哈尔滨工业大学计算机科学与技术学院，黑龙江哈尔滨 150001

修回日期:2021-11-02 出版日期:2021-11-25 发布日期:2021-11-01
作者简介:李骜（1986− ），男，黑龙江哈尔滨人，博士，哈尔滨理工大学副教授、博士生导师，主要研究方向为模式识别、机器学习等
王卓（1997− ），女，黑龙江牡丹江人，哈尔滨理工大学硕士生，主要研究方向为数据挖掘、机器学习
于晓洋（1962− ），男，黑龙江哈尔滨人，博士，哈尔滨理工大学教授、博士生导师，主要研究方向为三维视觉检测、图像处理等
陈德运（1962− ），男，黑龙江哈尔滨人，博士，哈尔滨理工大学教授、博士生导师，主要研究方向为探测与成像技术、图像处理等
张英涛（1975− ），女，黑龙江哈尔滨人，博士，哈尔滨工业大学副教授，主要研究方向为人工智能与信息处理等
孙广路（1979− ），男，黑龙江哈尔滨人，博士，哈尔滨理工大学教授、博士生导师，主要研究方向为机器学习、网络安全等
基金资助:
国家自然科学基金资助项目(62071157);黑龙江省自然科学基金资助项目(YQ2019F011);黑龙江省青年创新人才计划基金资助项目(UNPYSCT-2018203);黑龙江省高等学校基本科研业务费专项资金资助项目(LGYC2018JQ013);黑龙江省博士后启动基金资助项目(LBH-Q19112)

Robust multiview subspace clustering method based on multi-kernel low-redundancy representation learning

Ao LI¹^,², Zhuo WANG¹, Xiaoyang YU², Deyun CHEN¹, Yingtao ZHANG³, Guanglu SUN¹

¹ School of Computer Science and Technology, Harbin University of Science and Technology, Harbin 150080, China
² Postdoctoral Station of Instrument Science and Technology, Harbin University of Science and Technology, Harbin 150080, China
³ School of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China

Revised:2021-11-02 Online:2021-11-25 Published:2021-11-01
Supported by:
The National Natural Science Foundation of China(62071157);The Natural Science Foundation of Heilongjiang Province(YQ2019F011);University Nursing Program for Young Scholars with Creative Talents in Heilongjiang Province(UNPYSCT-2018203);The Fundamental Research Founds for the University of Heilongjiang Province(LGYC2018JQ013);Postdoctoral Foundation of Heilongjiang Province(LBH-Q19112)

摘要/Abstract

摘要：

针对高维数据冗余性、噪声干扰等问题对多视图子空间聚类性能的影响，提出一种多核低冗余表示学习的稳健多视图子空间聚类方法。首先，通过分析揭示数据在核空间中的冗余性和噪声影响特性，提出采用多核学习来获得局部视图数据的稳健低冗余表示，并利用其替代原始数据实施子空间学习。其次，引入张量分析模型进行多视图融合，从全局角度学习不同视图子空间表示的潜在张量低秩结构，在捕获视图间高阶相关性的同时保持其各异性专属信息。所提方法将稳健低冗余表示学习、视图专属子空间学习以及融合潜在子空间结构学习统一到一个目标函数中，使其在迭代中相互促进。大量实验结果表明，所提方法在多个客观评价指标方面均优于当前主流多视图聚类方法。

关键词: 低冗余表示学习, 子空间聚类, 多视图学习, 张量分析

Abstract:

Considering the impact of high dimensional data redundancy and noise interference on multiview subspace clustering, a robust multiview subspace clustering method based on multi-kernel low redundancy representation learning was proposed.Firstly, by analyzing and revealing the redundancy and noise influence characteristics of data in kernel space, a multi-kernel learning method was proposed to obtain a robust low-redundancy representation of local view-specific data, which was utilized to replace the original data to implement subspace learning.Secondly, a tensor analysis model was introduced to carry out multiview fusion, so as to learn the potential low-rank tensor structure among different subspace representations from global perspective.It would capture the high-order correlation among views while maintaining their unique information.In this method, robust low-redundancy representation learning, view-specific subspace learning and fusion potential subspace structure learning were unified into the same objective function, so that they could promote each other during iterations.A large number of experimental results demonstrate that the proposed method is superior to the existing mainstream multiview clustering methods on several objective evaluation indicators.

Key words: low-redundancy representation learning, subspace clustering, multiview learning, tensor analysis

中图分类号:

TP391

李骜, 王卓, 于晓洋, 陈德运, 张英涛, 孙广路. 多核低冗余表示学习的稳健多视图子空间聚类方法[J]. 通信学报, 2021, 42(11): 193-204.

Ao LI, Zhuo WANG, Xiaoyang YU, Deyun CHEN, Yingtao ZHANG, Guanglu SUN. Robust multiview subspace clustering method based on multi-kernel low-redundancy representation learning[J]. Journal on Communications, 2021, 42(11): 193-204.

图/表 8

图1

图2

表1

在BBC-Sport数据集上进行不同方法的ACC、NMI和F-measure比较"

方法	ACC	NMI	F-measure
LMSC	0.915±0.00	0.839±0.00	0.900±0.00
MSSC	0.971±0.00	0.898±0.00	0.944±0.00
PMSC	0.472±0.04	0.202±0.08	0.445±0.04
FMR	0.946±0.01	0.832±0.02	0.898±0.02
CoMSC	0.963±0.00	0.881±0.00	0.923±0.00
本文方法	$0 . 980 \pm 0 . 00$	$0 . 946 \pm 0 . 00$	$0 . 973 \pm 0 . 00$

表1

表2

在ORL数据集上进行不同方法的ACC、NMI和F-measure比较"

方法	ACC	NMI	F-measure
LMSC	0.811±0.03	0.908±0.02	0.747±0.04
MSSC	0.832±0.02	$0 . 927 \pm 0 . 01$	0.787±0.02
PMSC	0.551±0.04	0.732±0.03	0.394±0.06
FMR	0.809±0.02	0.898±0.01	0.741±0.02
CoMSC	0.812±0.01	0.901±0.00	0.749±0.01
本文方法	$0 . 846 \pm 0 . 03$	0.923±0.01	$0 . 791 \pm 0 . 02$

表2

表3

在UCI-Digits数据集上进行不同方法的ACC、NMI和F-measure比较"

方法	ACC	NMI	F-measure
LMSC	0.763±0.05	0.718±0.02	0.682±0.04
MSSC	0.924±0.00	0.859±0.00	0.858±0.00
PMSC	0.636±0.07	0.624±0.04	0.576±0.04
FMR	0.865±0.01	0.771±0.01	0.760±0.01
CoMSC	0.930±0.00	0.870±0.00	0.866±0.00
本文方法	$0 . 974 \pm 0 . 00$	$0 . 943 \pm 0 . 00$	$0 . 949 \pm 0 . 00$

表3

图3

图4

图5

参考文献 20

[1]	ZHAO J , XIE X J , XU X ,et al. Multi-view learning overview:recent progress and new challenges[J]. Information Fusion, 2017,38: 43-54.
[2]	ZHANG C Q , HU Q H , FU H Z ,et al. Latent multi-view subspace clustering[C]// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2017: 4333-4341.
[3]	ELHAMIFAR E , VIDAL R . Sparse subspace clustering:algorithm,theory,and applications[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013,35(11): 2765-2781.
[4]	LIU G C , LIN Z C , YAN S C ,et al. Robust recovery of subspace structures by low-rank representation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013,35(1): 171-184.
[5]	FU Y F , GAO J B , TIEN D ,et al. Tensor LRR and sparse coding-based subspace clustering[J]. IEEE Transactions on Neural Networks and Learning Systems, 2016,27(10): 2120-2133.
[6]	WANG J , WANG X , TIAN F ,et al. Constrained low-rank representation for robust subspace clustering[J]. IEEE Transactions on Cybernetics, 2017,47(12): 4534-4546.
[7]	YANG J F , LIANG J , WANG K ,et al. Subspace clustering via good neighbors[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020,42(6): 1537-1544.
[8]	XU C , TAO D , XU C . A survey on multi-view learning[J]. arXiv Preprint,arXiv:1304.5634, 2013.
[9]	CHAUDHURI K , KAKADE S M , LIVESCU K ,et al. Multi-view clustering via canonical correlation analysis[C]// Proceedings of Proceedings of the 26th Annual International Conference on Machine Learning. New York:ACM Press, 2009: 129-136.
[10]	GAO H C , NIE F P , LI X L ,et al. Multi-view subspace clustering[C]// Proceedings of 2015 IEEE International Conference on Computer Vision (ICCV). Piscataway:IEEE Press, 2015: 4238-4246.
[11]	LI R H , ZHANG C Q , HU Q H ,et al. Flexible multi-view representation learning for subspace clustering[C]// Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence.[S.n.:s.l], 2019: 2916-2922.
[12]	张茁涵, 曹容玮, 李晨 ,等. 隐式低秩稀疏表示的多视角子空间聚类[J]. 模式识别与人工智能, 2020,33(4): 344-352.
	ZHANG Z H , CAO R W , LI C ,et al. Latent low-rank sparse multi-view subspace clustering[J]. Pattern Recognition and Artificial Intelligence, 2020,33(4): 344-352.
[13]	ZHANG C Q , FU H Z , LIU S ,et al. Low-rank tensor constrained multiview subspace clustering[C]// Proceedings of 2015 IEEE International Conference on Computer Vision (ICCV). Piscataway:IEEE Press, 2015: 1582-1590.
[14]	ABAVISANI M , PATEL V M . Multimodal sparse and low-rank subspace clustering[J]. Information Fusion, 2018,39: 168-177.
[15]	KANG Z , ZHAO X J , PENG C ,et al. Partition level multiview subspace clustering[J]. Neural Networks, 2020,122: 279-288.
[16]	LIU J Y , LIU X W , YANG Y X ,et al. Multiview subspace clustering via co-training robust data representation[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021,PP(99): 1-13.
[17]	CAO X C , ZHANG C Q , FU H Z ,et al. Diversity-induced Multi-view subspace clustering[C]// Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2015: 586-594.
[18]	SCH?LKOPF B , SMOLA A , MüLLER K R , . Nonlinear component analysis as a kernel eigenvalue problem[J]. Neural Computation, 1998,10(5): 1299-1319.
[19]	BOYD S . Distributed optimization and statistical learning via the alternating direction method of multipliers[J]. Foundations and Trends? in Machine Learning, 2010,3(1): 1-122.
[20]	KILMER M E , BRAMAN K , HAO N ,et al. Third-order tensors as operators on matrices:a theoretical and computational framework with applications in imaging[J]. SIAM Journal on Matrix Analysis and Applications, 2013,34(1): 148-172.

多核低冗余表示学习的稳健多视图子空间聚类方法

Robust multiview subspace clustering method based on multi-kernel low-redundancy representation learning

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 20

相关文章 2

Metrics

推荐阅读 0

[1]	魏晶晶,陈畅,廖祥文,陈国龙,程学旗. 基于受限非负张量分解的用户社会影响力分析[J]. 通信学报, 2016, 37(6): 154-162.
[2]	朱玉娜,韩继红,袁霖,范钰丹,陈韩托,谷文. 面向密码协议在线安全性的监测方法[J]. 通信学报, 2016, 37(6): 75-85.