Local Structure and High-Order Feature Preserved Network Embedding Based on Non-Negative Matrix Factorization

被引:0
|
作者
Tian, Qin [1 ]
Pan, Lin [1 ]
Guo, Xuan [2 ]
Li, Xiaoming [3 ]
Yu, Wei [3 ]
Li, Faming [4 ]
机构
[1] Tianjin Univ, Sch Marine Sci & Technol, Tianjin 300072, Peoples R China
[2] Tianjin Univ, Coll Intelligence & Comp, Tianjin 300350, Peoples R China
[3] Zhejiang Yuexiu Univ, Sch Int Business, Shaoxing 312069, Peoples R China
[4] Res Inst Chem Def, Beijing 102205, Peoples R China
来源
IEEE ACCESS | 2020年 / 8卷
基金
中国国家自然科学基金;
关键词
Task analysis; Matrix decomposition; Sparse matrices; Topology; Feature extraction; Natural language processing; Licenses; Network embedding; local structure; high-order feature; non-negative matrix factorization; LINK-PREDICTION; SOCIAL NETWORKS; EQUIVALENCE;
D O I
10.1109/ACCESS.2020.3045532
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Network embedding, as an effective method of learning the low-dimensional representations of nodes, has been widely applied to various complex network analysis tasks, such as node classification, community detection, link prediction and evolution analysis. The existing embedding methods usually focus on the local structure of the network by capturing community structure, first-order or second-order proximity, etc. Some methods have been proposed to model the high-order proximity of networks to capture more effective information. However, they are incapable of preserving the similarity among nodes that are not very close to each other in network but have similar structures. For instance, the nodes with similar local topology structure should be similar in embedding space even if they are not in the same community. Herein, we regard these structure characteristics as the high-order features, which reveals that the structure similarity between nodes is spatially unrelated. In light of above the limitations of existing methods, we construct the high-order feature matrix for mutually reinforcing the embedding which preserves the local structure. To integrate these features effectively, we propose LHO-NMF, which fuses the high-order features into non-negative matrix factorization framework while capturing the local structure. The proposed LHO-NMF could effectively learn the node representations via preserving the local structure and high-order feature information. In specific, the high-order features are learned based on random walk algorithm. The experimental results show that the proposed LHO-NMF method is very effective and outperforms other state-of-the-art methods among multiple downstream tasks.
引用
收藏
页码:225967 / 225980
页数:14
相关论文
共 50 条
  • [1] Lower order information preserved network embedding based on non-negative matrix decomposition
    Tian, Qiang
    Pan, Lin
    Zhang, Wang
    Li, Tianpeng
    Wu, Huaming
    Jiao, Pengfei
    Wang, Wenjun
    INFORMATION SCIENCES, 2021, 572 : 43 - 56
  • [2] Feature Weighted Non-Negative Matrix Factorization
    Chen, Mulin
    Gong, Maoguo
    Li, Xuelong
    IEEE TRANSACTIONS ON CYBERNETICS, 2023, 53 (02) : 1093 - 1105
  • [3] A New Feature Extraction and Recognition Method for Microexpression Based on Local Non-negative Matrix Factorization
    Gao, Junli
    Chen, Huajun
    Zhang, Xiaohua
    Guo, Jing
    Liang, Wenyu
    FRONTIERS IN NEUROROBOTICS, 2020, 14
  • [4] Image Feature Extraction via Graph Embedding Regularized Projective Non-negative Matrix Factorization
    Du, Haishun
    Hu, Qingpu
    Zhang, Xudong
    Hou, Yandong
    PATTERN RECOGNITION (CCPR 2014), PT I, 2014, 483 : 196 - 209
  • [5] Non-negative Matrix Factorization Meets Word Embedding
    Ailem, Melissa
    Salah, Aghiles
    Nadif, Mohamed
    SIGIR'17: PROCEEDINGS OF THE 40TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL, 2017, : 1081 - 1084
  • [6] Local Coordinate Projective Non-negative Matrix Factorization
    Liao, Qing
    Zhang, Qian
    Zhang, Xiang
    Guan, Naiyang
    2015 IEEE 14TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA), 2015, : 604 - 607
  • [7] Binary Codes Based on Non-Negative Matrix Factorization for Clustering and Retrieval
    Xiong, Jiang
    Tao, Yingyin
    Zhang, Meng
    Li, Huaqing
    IEEE ACCESS, 2020, 8 : 207012 - 207023
  • [8] Non-negative Matrix Factorization For Network Delay Matrix Completion
    Ghandi, Sanaa
    Reiffers-Masson, Alexandre
    Vaton, Sandrine
    Chonavel, Thierry
    PROCEEDINGS OF THE IEEE/IFIP NETWORK OPERATIONS AND MANAGEMENT SYMPOSIUM 2022, 2022,
  • [9] FEATURE EXTRACTION VIA MULTI-VIEW NON-NEGATIVE MATRIX FACTORIZATION WITH LOCAL GRAPH REGULARIZATION
    Wang, Zhenfan
    Kong, Xiangwei
    Fu, Haiyan
    Li, Ming
    Zhang, Yujia
    2015 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2015, : 3500 - 3504
  • [10] FULLY SUPERVISED NON-NEGATIVE MATRIX FACTORIZATION FOR FEATURE EXTRACTION
    Austin, Woody
    Anderson, Dylan
    Ghosh, Joydeep
    IGARSS 2018 - 2018 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM, 2018, : 5772 - 5775