Statistical inference in complex networks

Abstract

The epidemic and rumor spreading can be naturally modeled as a dynamical process in complex networks, where vertices represent individuals and their connections and interactions are represented by edges. Previous studies have shown that the network structure, that is, the way that nodes connect influence the spreading processes. The available real networks data are samples of the complete network. Therefore, it is interesting to study the effects of the existing sampling methods and which are the most influent topological features in the spreading processes.In this project, we propose to study the existing sampling techniques and to use regression analysis in order to discover the most important topological properties in the epidemic and rumor spreading. Besides, we intend to predict the quantity of infected or informant nodes throughout the spreading using the generated regression model.