Identifying influential nodes in complex networks via weighted k-shell entropy-based approach

Identifying nodes that exert significant influence within complex networks presents a considerable challenge, primarily due to the vast number of nodes involved. Most current detection techniques utilize either the degree or topological position of nodes to identify those that may be affected. However, these methods mainly concentrate on utilizing either local or global information derived from the network, which can lead to imprecise detection outcomes. To address this issue, this paper presents a technique for detecting influential nodes using weighted $k$-shell entropy. In this framework, the influence of a node is determined by the entropy obtained from its local and global information. The $k$-shell value and clustering coefficient are employed to quantify global information entropy, while the node’s degree represents local information. Additionally, the method accounts for the influence exerted between nodes and their neighbors through a weighting mechanism. Furthermore, the paper examines the impact ranking problem as it pertains to the susceptible–infection–recovery model and evaluates the proposed ranking method in this context. Experimental findings on both synthetic random networks and real-world networks indicate that the proposed method attains higher accuracy than other current techniques.