Semi-global triangular centrality measure for identifying the influential spreaders from undirected complex networks

The influential spreaders play a significant role in maximizing or controlling any spreading process in a network. In the literature, many methods have been proposed to identify influential spreaders. In this article, we classify all the methods mainly into four categories, such as local centrality, global centrality, semi-global centrality and hybrid centrality. Among them, we have found semi-global centrality based methods have immense potential in identifying the influential spreaders from various types of network structures. However, we have observed that the existing semi-global centrality methods can identify the spreaders from the periphery of a network, where the nodes in the periphery are loosely coupled and the collective influence in the peripheral region of a spreading process will be nominal. We propose a new indexing method semiglobal triangular centrality, which does not consider the best spreaders from the periphery. The proposed method maximizes the total collective influence of a spreading process by selecting the best spreaders from the dense part of a network. We have examined the performance of the proposed method using the Susceptible-Infected-Recovered epidemic model and applied to nine real-networks. The experimental result reveals that the proposed method performs better than the other centrality methods in terms of spreading dynamics.

Automated summary

This article classifies the methods for identifying influential spreaders into four categories: local centrality, global centrality, semi-global centrality, and hybrid centrality. Semi-global centrality based methods are found to be highly effective in identifying influential spreaders from different network structures. However, existing methods tend to overlook the spreaders from the periphery of a network. To address this issue, the authors propose a new indexing method called semiglobal triangular centrality, which focuses on selecting the best spreaders from the dense part of a network. Experimental results show that this method outperforms other centrality methods in terms of spreading dynamics.