Blocking Complex Contagions Using Community Structure

Blocking the propagation of contagions in populations has many applications, such as stopping the dissemination of leaked information and impeding the spread of an ideology or opinion. Several methods exist for blocking contagions by removing key or critical nodes from a network representation of a population. Some methods, based on network topology alone, run very fast, but their performance can be inferior to slower-running blocking methods that also take into account contagion dynamics. Further, most works focus on simple contagions, where a node can contract a contagion from one neighbor. In contrast, complex contagions propagate via interactions with two or more neighbors possessing a contagion. Little work has been done on blocking complex contagions. We present an approach for selecting critical nodes for simple and complex contagions. In selecting critical nodes, our hybrid method uses topology-only information and thereafter couples those results with information about contagion dynamics. We evaluate our method using three different kinds of social networks from the literature.