Social Network Analysis for the Identification of Key Spreaders During COVID-19

Abstract

Identifying key spreaders is regarded as one of the fundamental challenging areas in controlling the spread of infections caused due to deadly Coronavirus Disease 2019 (COVID-19). Identification of key spreaders greatly contributes towards the understanding of disease spreading mechanisms during pandemic. The resolution of this problem remains very useful for government agencies to make tactical and actionable plans to counter the rapid spread of COVID-19. Various researchers around the world are adopting computational techniques to recognize key spreaders efficiently. However, the development of such technologies poses a significant challenge and requires regular improvements over time. In this chapter, a method CovidKeySpreader has been proposed that integrates both local and global insight and adopts a random walk algorithm to determine key spreaders in a patient interaction network related to India’s different states. For implementation, a state-wise network is constructed where nodes represent an individual patient and edges show interaction link between them. Each node of the network is assigned to associated communities. Further, communities are analyzed by exploiting both node and community scores. Finally, a random walk algorithm is applied to the weighted network to iteratively rank nodes. The efficacy of the proposed method is established using Susceptible-Infected-Recovered (SIR) spreading model and simulate the process of spreading on networks. Experiments conducted on four state-wise networks. The evaluation metric shows that the key spreaders identified by our proposed algorithm are more significant in comparison to other basic centrality measures.