Community deception in directed influence networks

Community deception is about protecting users of a community from being discovered by community detection algorithms. This paper studies community deception in directed influence network (DIN). It aims to address the limitations of the state of the art through a twofold strategy: introducing directed influence and considering the role of nodes in the deception strategy. The study focuses on using modularity as the optimization function. It offers several contributions, including an upgraded version of modularity that accommodates the concept of influence, edge-based, and node-based deception algorithms. The study concludes with a comparison of the proposed methods with the state of the art showing that not only influence is a valuable ingredient to devising deception strategies but also that novel deception approaches centered on node operations can be successfully devised.