The collaborative role of K-Shell and PageRank for identifying influential nodes in complex networks

Abstract

Finding the most influential nodes in complex networks is a significant challenge with applications in various fields, including social networks, biology, and transportation systems. Many existing methods rely on different structural properties but often overlook complementary features. This paper highlights the complementary nature of K-Shell and PageRank and proposes a novel linear metric that combines them. Through extensive comparisons of 19 real-world and several artificial networks, the proposed method demonstrates superior accuracy, resolution, and computational efficiency. Evaluations against 11 state-of-the-art methods, including IDME, HGSM, and DNC, underscore the superiority of the proposed approach. Notably, the average accuracy has increased by 33.3% compared to PageRank and 23.1% compared to K-Shell, emphasizing the importance of integrating these two features.