A novel potential edge weight method for identifying influential nodes in complex networks based on neighborhood and position

The identification of influential nodes is one of the most urgent and challenging research issues in complex networks, which is crucial to the robustness and stability of networks. Edges in networks are vital communication channels in the information spreading process, many existing methods for identifying influential nodes are established via employing characteristics of nodes, but most of them consider the edges equally in unweighted networks. In this paper, we introduce a novel potential edge weight to distinguish the influence of each edge by utilizing the H-index, k-shell iteration factor and clustering coefficient of its connection nodes. A novel potential edge weight method, called the HIC centrality (HIC), is proposed to identify influential nodes in complex networks. The key idea of HIC is to incorporate neighborhood, position, and topological structure features to identify influential nodes. To evaluate the performance of HIC, we adopt the susceptible–infected–recovered (SIR) model and the susceptible–infected (SI) model for simulation and conduct a series of comparative experiments on nine real-world networks and three artificial networks. Experimental results show that the proposed method is superior to seven contrast methods in terms of accuracy, effectiveness and distinguishing ability, including five centrality measures and two popular potential edge weight methods. The HIC centrality has low computational complexity, allowing it to be applied to sparse large-scale networks. •A novel potential edge weight method is proposed to identify influential nodes in complex networks.•The proposed HIC method assigns different weight to each edge of networks based on neighborhood, position and topology structure characteristics.•HIC outperforms seven state-of-the-art methods in terms of accuracy, effectiveness and distinguishing ability.•HIC has low computational complexity and it can be applied to sparse large-scale networks.