An Efficient Immunization Strategy Using Overlapping Nodes and Its Neighborhoods

When an epidemic occurs, it is often impossible to vaccinate the entire population due to limited amount of resources. Therefore, it is of prime interest to identify the set of influential spreaders to immunize, in order to minimize both the cost of vaccine resource and the disease spreading. While various strategies based on the network topology have been introduced, few works consider the influence of the community structure in the epidemic spreading process. Nowadays, it is clear that many real-world networks exhibit an overlapping community structure, in which nodes are allowed to belong to more than one community. Previous work shows that the numbers of communities to which a node belongs is a good measure of its epidemic influence. In this work, we address the effect of nodes in the neighborhood of the overlapping nodes on epidemics spreading. The proposed immunization strategy provides highly connected neighbors of overlapping nodes in the network to immunize. The whole process requires information only at the node level and is well suited to large-scale networks. Extensive experiments on four real-world networks of diverse nature have been performed. Comparisons with alternative local immunization strategies using the fraction of the Largest Connected Component (LCC) after immunization,show that the proposed method is much more efficient. Additionally, it compares favorably to global measures such as degree and betweenness centrality.