VIKOR based Methodology using Local and Global Centralities to Detect Spreaders in Complex Network

Abstract

The identification of node spreaders is of great importance for influence maximization in complex networks. Nodes that have a high degree, high eigenvector, high betweenness, and high closeness have been identified as spreaders in previous research. Centralities that can be computed using local information of the node have low time complexity but don’t consider the whole network. Centralities that use global information of the network can’t be applied to large-scale networks and have more time complexity although their high accuracy. In this paper, we propose a novel integrated methodology that combines local and global centralities with the MCDM technique “VIKOR” to synthesize the spreaders nodes. To validate the proposed methodology, we test it on real networks, and the obtained results are satisfactory and prove that it behaves well to find the spreaders nodes in these networks.