| Frontiers in Physics | |
| Identifying vital nodes in recovering dynamical process of networked system | |
| Physics | |
| Xiaoya Jiang1  Jiale Fu1  Qi Shao1  Wenwu Yu2  Duxin Chen3  | |
| [1] Jiangsu Key Laboratory of Networked Collective Intelligence, School of Mathematics, Southeast University, Nanjing, China;null;wwyu@seu.edu.cn; | |
| 关键词: vital nodes identification; complex networks; deep learning; graph convolutional network; network recovery; | |
| DOI : 10.3389/fphy.2023.1167585 | |
| received in 2023-02-16, accepted in 2023-05-25, 发布年份 2023 | |
| 来源: Frontiers | |
PDF
|
|
【 摘 要 】
Vital nodes identification is the problem of identifying the most significant nodes in complex networks, which is crucial in understanding the property of the networks and has applications in various fields such as pandemic controlling and energy saving. Traditional methods mainly focus on some types of centrality indices, which have restricted application cases. To improve the flexibility of the process and enable simultaneous multiple nodes mining, a deep learning-based vital nodes identification algorithm is proposed in this study, where we train the influence score of each node by using a set of nodes to approximate the rest of the network via the graph convolutional network. Experiments are conducted with generated data to justify the effectiveness of the proposed algorithm. The experimental results show that the proposed method outperforms the traditional ways in adaptability and accuracy to recover the dynamical process of networked system under different classes of network structure.
【 授权许可】
Unknown
Copyright © 2023 Fu, Jiang, Shao, Chen and Yu.
【 预 览 】
| Files | Size | Format | View |
|---|---|---|---|
| RO202310100324787ZK.pdf | 1745KB |
PDF