Identifying Influential Nodes in Complex System from Multiscale Complex Network Perspective

Modeling complex systems as networks and identifying influential nodes is still an open and challenging issue due to the increasing system scale, complex structure, and dynamic behaviors. A number of methods have been recently proposed to solve this problem but most existing studies have the limitations, and few of them considering multiscale information of the network. In this article, therefore, a novel method to identify influential nodes is proposed, which takes into account three scales attributes, i.e., local, semi-local and global, instead of only a single scale structure attribute. Empirical studies were performed on six Open-source software projects, and the experimental results indicated that the proposed method outperformed traditional baseline approaches. Our findings demonstrate the proposed method is an effective method for identifying influential nodes in complex system, and this study is expected to provide useful insights into smart control of complex system.