Improving the identification of influential nodes in complex networks using semi-local structures and shortest-path-based centralities

Identifying influential nodes in complex networks is crucial for analyzing network structure, information dissemination, and controlling network dynamics. Traditional centrality-based methods typically focus on individual topological features and often overlook the combined influence of nodes and their interactions across multiple levels of the network. Recent advances in semi-local centralities have improved performance; however, balancing accuracy and computational complexity in large-scale networks remains a challenge. To address this limitation, this paper proposes a shortest-path-based semi-local centrality measure that operates on semi-local structures (SPSLS). SPSLS integrates several shortest-path-based features, including shortest path lengths, number of shortest paths, and average shortest paths, to better capture the topological influence exerted by neighboring nodes. To enhance scalability, SPSLS adopts an extended neighborhood strategy within a distributed framework, considering only relevant portions of the network to quantify each node’s influence efficiently. Consequently, SPSLS simultaneously accounts for semi-local structural information and network path features, achieving a balance between performance and complexity in node centrality analysis. The effectiveness of SPSLS is evaluated on real-world networks using the susceptible-infected-recovered (SIR) model and Kendall’s tau correlation, showing improved performance in information propagation compared to existing centrality measures. Specifically, SPSLS outperforms the best existing measure by 1.2%.