Identifying Important Nodes in Multi-Relational Networks Based on Evidence Theory

被引：0
|
作者：
Luo H. [1 ]
Yan G.-H. [1 ]
Zhang M. [1 ]
Bao J.-B. [1 ]
Li J.-C. [1 ]
Liu T. [1 ]
机构：
[1] School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou
来源：
Yan, Guang-Hui (yanghacademic@163.com) | 1600年 / Science Press卷 / 43期
基金：
中国国家自然科学基金;
关键词：
Centrality; Evidence theory; Important nodes; Local clustering coefficient; Multi-relational network; Multiplex network; Transitivity;
D O I：
10.11897/SP.J.1016.2020.02398
中图分类号：
TP [自动化技术、计算机技术];
学科分类号：
0812 ;
摘要：
As a typical model of the real world, the multi-relational network has become one of hot topics in the field of network science. Identifying important nodes is one of the fundamental problem of network analysis, which is essential to understand the structure and dynamic characteristics of the complex networks. Lots of excellent studies carried on the importance of nodes in a single-relational network, but the research on important nodes in multi-relational networks lacks systematic research. Trying to explore and extend the existing methods in single networks to multi-relational networks has become one of the hot issues. For multi-relational networks, this paper focused on the study of the identifying importance nodes considering the influence of centrality and transitivity. We created the multilayer network model to describe a multi-relational network and used the multilayer adjacency matrix to represent it. To quantify the transitivity in multi-relational networks, we proposed a method for calculating the local clustering coefficient according to the notion of clustering coefficient for multilayer networks. Ulteriorly, we proposed multiplex ClusterRank combining the multiplex degree centrality and extending ClusterRank in single-relational networks to multi-relational networks. Considering the influence of coupling information and the difference of transmission mechanism for multi-relational network on vital information of nodes, multiplex ClusterRank may have calculation errors in large-scale multi-relational networks, which may affect the accuracy of measuring essential nodes. Therefore, in this work we improved multiplex ClusterRank and proposed another more efficient method to identify vital nodes for multi-relational networks named multiplex evidential centrality, which fused the node degree centrality and local clustering coefficient by the Dempster-Shafer evidence theory. Numerous experiments were carried out on four real networks. The proposed method is evaluated from two aspects: robustness and vulnerability, transmission dynamics characteristics. Experimental results show that multiplex evidential centrality can get better results than multiplex degree centrality and multiplex eigenvector centrality which only focus on the structure of multiplex networks and multiplex ClusterRank does have calculation errors in large-scale networks, making it difficult to identify important nodes. The experimental results verified multiplex evidential centrality can effectively eliminate the influence of the coupling information and the transmission mechanism in multi-relational networks, and has lower time complexity than multiplex ClusterRank. In this paper, we not only provide new ideas and methods for identifying critical nodes for multi-relational networks but also expands the application of information fusion technology. © 2020, Science Press. All right reserved.
引用
收藏
页码：2398 / 2413
页数：15
相关论文
共 41 条
  • [1] Cohen R, Erez K, Benavraham D, Et al., Breakdown of the Internet under intentional attack, Physical Review Letters, 86, 16, pp. 3682-3685, (2001)
  • [2] Albert R, Jeong H, Barabasi A., Error and attack tolerance of complex networks, Nature, 406, pp. 378-382, (2001)
  • [3] Fang Jin-Qing, From a single network to "network of networks" development process: Some discussions on the exploration of multilayer supernetwork models and challenges, Complex Systems and Complexity Science, 13, 1, pp. 40-47, (2016)
  • [4] Wang Na-Na, Gao Hong, Liu Wei, Research on a heterogeneous edge multi-graph network model, CAAI Transactions on Intelligent Systems, 12, 4, pp. 475-481, (2017)
  • [5] Wang Na-Na, Gao Hong, Li Shan-Shan, Et al., A research on hypergraph of heterogeneous edge, Journal of Guangdong University of Technology, 34, 1, pp. 6-10, (2017)
  • [6] Huang Rui-Yang, Wu Qi, Zhu Yu-Hang, Community detection in dynamic heterogeneous network with joint nonnegative matrix factorization, Application Research of Computers, 34, 10, pp. 2989-2992, (2017)
  • [7] Bin Sheng, Sun Geng-Xin, Important node detection algorithm for multiple relationships on line social network based on multi-subnet composited complex network model, Journal of Nanjing University: Natural Sciences, 53, 2, pp. 378-385, (2017)
  • [8] Kolda T G, Bader B W., Tensor decompositions and applications, SIAM Review, 51, 3, pp. 455-500, (2009)
  • [9] Jo H, Moon H, Ki Baek S., Immunization dynamics on a 2-layer network model, Physica A: Statistical Mechanics and Its Applications, 361, 2, pp. 534-542, (2006)
  • [10] Acar E, Yener B., Unsupervised multiway data analysis: A literature survey, IEEE Transactions on Knowledge and Data Engineering, 21, 1, pp. 6-20, (2009)
12345