Computer Virus Propagation in Social Networks

This paper applies social network analysis techniques to study computer virus propagation. We propose a novel multilevel hierarchical linear model to simultaneously evaluate the impact of both individual-level and group-level variables on virus propagation process. In this model, we propose centrality and brokerage measures as explanatory variables. We estimate our model based on empirical data from the largest social networking website and find that closeness centrality (individual level) and brokerage (group level) jointly explain 82.5% of the variance in the number of infections. This research contributes to both the literature of computer virus propagation and defense and the literature of centrality measure comparison in the filed of social network analysis by: (1) performing subgroup analysis and considering multiple levels of network characteristics to capture the intrinsic nested feature of the networks and (2) comparing different structural measures (centrality and brokerage) in terms of their performance to explain the propagation of computer viruses.