.png){kind=icon}
Return
A key node identification method based on neighborhood-derived cluster method
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2026.134646.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Identifying key nodes in complex networks has been a fundamental and challenging task within the field of net work science, including information dissemination, virus containment, public opinion monitoring, and so on. However, conventional methodologies predominantly rely on the local or global characteristics of the network topology, which inadequately capture the polyadic interactions inherent to numerous real-world systems. In this paper, we propose a neighborhood-derived cluster method, termed NDCM, a method that integrates clustering structure with multi-dimensional topological features. Our approach first quantifies the clustering potential of each node and computes the membership score to detect community structure. Subsequently, diffusion capabil ity is introduced to select a global hub node within each community to ensure efficient information propagation across communities. Thereafter, an exponential decay-weighted mechanism integrates multidimensional topo logical features, including neighborhoods, positions, structural redundancy, etc., to obtain the comprehensive influence of nodes. Extensive experiments are verified on nine real-world networks, demonstrating NDCM's outstanding performance compared to benchmarks across multiple metrics, incorporating Kendall's tau coeffi cient, imprecision function, monotonicity, and complementary cumulative distribution function (CCDF). Notably, NDCM manifests a superiority over existing centrality methods in terms of predictive accuracy of node influence, sorting stability, and robustness.
Keywords:
Complex networks
Key node identification
Clustering algorithm
Multi-dimensional topology
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W
Organization
Cited Papers
No cited papers available