Wei Lin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Wei Lin has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Li Xu
College of Computer and Cyber Security, Fujian Provincial Key Laboratory of Network Security and Cryptology, Fujian Normal University, Fuzhou, China
4 shared publications - Kun Xie
College of Computer Science and Electronics Engineering, Hunan University, Changsha, China
2 shared publications - Min Li
Fujian Normal University
2 shared publications - Shuming Zhou
College of Mathematics and Statistics, Fujian Normal University, Fuzhou, China
2 shared publications - Chen Li
College of Computer and Cyber Security and Fujian Provincial Key Laboratory of Network Security and Cryptology, Fujian Normal University, Fuzhou, China
1 shared publication - Gaogang Xie
Chinese Academy of Sciences
1 shared publication
Latest Publications
- Open Access
- 23 September 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 298
Interlink reconfiguration against cascading failures on cyber-physical power systems based on an improved memetic algorithm
- 23 September 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 298
Open AccessInterlink reconfiguration against cascading failures on cyber-physical power systems based on an improved memetic algorithm
- 25 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Services Computing
- Vol. 17 (6) , 3597-3612
Enhanced Network Traffic Anomaly Detection: Integration of Tensor Eigenvector Centrality with Low-Rank Recovery Models
In service computing, network traffic anomaly detection is pivotal for monitoring and identifying irregularities in network traffic to uphold the security, reliability, and stability of networks and services. In network traffic data, centrality is exhibited as certain nodes more frequently act as communication sources or destinations, or play critical intermediary roles in the network. These structures are also among the targets of network bottlenecks and targeted attacks. Current unsupervised network traffic anomaly detection algorithms, based on low-rank tensor recovery, achieve effective detection performance by comprehensively capturing network information. However, these algorithms often neglect the underlying topological structure, focusing solely on linear data structures, which leads to overlooking the degree of traffic concentration and nonlinear data structures. It reduces the detection efficiency of abnormal traffic generated by targeted attacks. To comprehensively understand the evolution of traffic concentration over time, this study introduces a mathematical formula for tensor eigenvector edge centrality. The formula provides rankings of edge importance based on the significance of nodes and time layers, and the effectiveness of centrality is validated through structural perturbations in the network. On this basis, we design a low-rank tensor recovery model utilizing representation learning to obtain the centrality feature matrix of network traffic data. By encoding centrality for nonlinear proximity information, and incorporating the Laplacian matrix to capture nonlinear structural information in tensor decomposition, the accuracy of anomaly detection is enhanced. Extensive experiments on Abilene and GÈANT network traffic data demonstrate that our proposed algorithm not only achieves higher precision and recall rates in random anomalies but also performs better in detecting anomalous traffic generated by high centrality structures compared to state of art algorithms based on matrix-based anomaly detection and tensor recovery methods.
- 25 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Services Computing
- Vol. 17 (6) , 3597-3612
Enhanced Network Traffic Anomaly Detection: Integration of Tensor Eigenvector Centrality with Low-Rank Recovery Models
In service computing, network traffic anomaly detection is pivotal for monitoring and identifying irregularities in network traffic to uphold the security, reliability, and stability of networks and services. In network traffic data, centrality is exhibited as certain nodes more frequently act as communication sources or destinations, or play critical intermediary roles in the network. These structures are also among the targets of network bottlenecks and targeted attacks. Current unsupervised network traffic anomaly detection algorithms, based on low-rank tensor recovery, achieve effective detection performance by comprehensively capturing network information. However, these algorithms often neglect the underlying topological structure, focusing solely on linear data structures, which leads to overlooking the degree of traffic concentration and nonlinear data structures. It reduces the detection efficiency of abnormal traffic generated by targeted attacks. To comprehensively understand the evolution of traffic concentration over time, this study introduces a mathematical formula for tensor eigenvector edge centrality. The formula provides rankings of edge importance based on the significance of nodes and time layers, and the effectiveness of centrality is validated through structural perturbations in the network. On this basis, we design a low-rank tensor recovery model utilizing representation learning to obtain the centrality feature matrix of network traffic data. By encoding centrality for nonlinear proximity information, and incorporating the Laplacian matrix to capture nonlinear structural information in tensor decomposition, the accuracy of anomaly detection is enhanced. Extensive experiments on Abilene and GÈANT network traffic data demonstrate that our proposed algorithm not only achieves higher precision and recall rates in random anomalies but also performs better in detecting anomalous traffic generated by high centrality structures compared to state of art algorithms based on matrix-based anomaly detection and tensor recovery methods.
- 1 October 2022
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 32 (10) , 103131
Finding influential edges in multilayer networks: Perspective from multilayer diffusion model
With the popularization of social network analysis, information diffusion models have a wide range of applications, such as viral marketing, publishing predictions, and social recommendations. The emergence of multiplex social networks has greatly enriched our daily life; meanwhile, identifying influential edges remains a significant challenge. The key problem lies that the edges of the same nodes are heterogeneous at different layers of the network. To solve this problem, we first develop a general information diffusion model based on the adjacency tensor for the multiplex network and show that the -mode singular value can control the level of information diffusion. Then, to explain the suppression of information diffusion through edge deletion, efficient edge eigenvector centrality is proposed to identify the influence of heterogeneous edges. The numerical results from synthetic networks and real-world multiplex networks show that the proposed strategy outperforms some existing edge centrality measures. We devise an experimental strategy to demonstrate that influential heterogeneous edges can be successfully identified by considering the network layer centrality, and the deletion of top edges can significantly reduce the diffusion range of information across multiplex networks.- 1 October 2022
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 32 (10) , 103131
Finding influential edges in multilayer networks: Perspective from multilayer diffusion model
With the popularization of social network analysis, information diffusion models have a wide range of applications, such as viral marketing, publishing predictions, and social recommendations. The emergence of multiplex social networks has greatly enriched our daily life; meanwhile, identifying influential edges remains a significant challenge. The key problem lies that the edges of the same nodes are heterogeneous at different layers of the network. To solve this problem, we first develop a general information diffusion model based on the adjacency tensor for the multiplex network and show that the -mode singular value can control the level of information diffusion. Then, to explain the suppression of information diffusion through edge deletion, efficient edge eigenvector centrality is proposed to identify the influence of heterogeneous edges. The numerical results from synthetic networks and real-world multiplex networks show that the proposed strategy outperforms some existing edge centrality measures. We devise an experimental strategy to demonstrate that influential heterogeneous edges can be successfully identified by considering the network layer centrality, and the deletion of top edges can significantly reduce the diffusion range of information across multiplex networks.- 9 April 2022
- Published by Elsevier in Journal of Computational Science
- Vol. 61
Identifying key nodes in interdependent networks based on Supra-Laplacian energy
- 9 April 2022
- Published by Elsevier in Journal of Computational Science
- Vol. 61
Identifying key nodes in interdependent networks based on Supra-Laplacian energy
- Open Access
- 29 June 2021
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2021, 1-10
The Hermitian Kirchhoff Index and Robustness of Mixed Graph
Large-scale social graph data poses significant challenges for social analytic tools to monitor and analyze social networks. The information-theoretic distance measure, namely, resistance distance, is a vital parameter for ranking influential nodes or community detection. The superiority of resistance distance and Kirchhoff index is that it can reflect the global properties of the graph fairly, and they are widely used in assessment of graph connectivity and robustness. There are various measures of network criticality which have been investigated for underlying networks, while little is known about the corresponding metrics for mixed networks. In this paper, we propose the positive walk algorithm to construct the Hermitian matrix for the mixed graph and then introduce the Hermitian resistance matrix and the Hermitian Kirchhoff index which are based on the eigenvalues and eigenvectors of the Hermitian Laplacian matrix. Meanwhile, we also propose a modified algorithm, the directed traversal algorithm, to select the edges whose removal will maximize the Hermitian Kirchhoff index in the general mixed graph. Finally, we compare the results with the algebraic connectivity to verify the superiority of the proposed strategy.
- 29 June 2021
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2021, 1-10
Open AccessThe Hermitian Kirchhoff Index and Robustness of Mixed Graph
Large-scale social graph data poses significant challenges for social analytic tools to monitor and analyze social networks. The information-theoretic distance measure, namely, resistance distance, is a vital parameter for ranking influential nodes or community detection. The superiority of resistance distance and Kirchhoff index is that it can reflect the global properties of the graph fairly, and they are widely used in assessment of graph connectivity and robustness. There are various measures of network criticality which have been investigated for underlying networks, while little is known about the corresponding metrics for mixed networks. In this paper, we propose the positive walk algorithm to construct the Hermitian matrix for the mixed graph and then introduce the Hermitian resistance matrix and the Hermitian Kirchhoff index which are based on the eigenvalues and eigenvectors of the Hermitian Laplacian matrix. Meanwhile, we also propose a modified algorithm, the directed traversal algorithm, to select the edges whose removal will maximize the Hermitian Kirchhoff index in the general mixed graph. Finally, we compare the results with the algebraic connectivity to verify the superiority of the proposed strategy.