Locating influential nodes via dynamics-sensitive centrality

TL;DR

This paper introduces a dynamics-sensitive centrality measure that combines network topology and dynamics to more accurately identify influential nodes in spreading processes, outperforming traditional centrality metrics.

Contribution

The paper presents a novel dynamics-sensitive centrality that improves the identification of influential spreaders by integrating topological and dynamical information.

Findings

DS centrality outperforms degree, k-shell, and eigenvector centrality in accuracy.

Empirical results on four real networks validate the effectiveness of DS centrality.

Applicable to both SIR and SI spreading models.

Abstract

With great theoretical and practical significance, locating influential nodes of complex networks is a promising issues. In this paper, we propose a dynamics-sensitive (DS) centrality that integrates topological features and dynamical properties. The DS centrality can be directly applied in locating influential spreaders. According to the empirical results on four real networks for both susceptible-infected-recovered (SIR) and susceptible-infected (SI) spreading models, the DS centrality is much more accurate than degree, $k$-shell index and eigenvector centrality.