Community Deception - Or: How to Stop Fearing Community Detection Algorithms (Extended Abstract)

Community deception is the problem of hiding a community from community detection algorithms. This is an important task whenever a group (e.g., activists, police enforcements) want to observe and cooperate in a social network while avoiding to be detected. We formalize the community deception problem and propose an efficient algorithm, based on the concept of community safeness, which allows to achieve deception by carefully identifying and rewiring a certain number of the community members' connections. Deception can be practically achieved in social networks like Facebook or Twitter by friending (following) or unfriending (unfollowing) network members. We validated our approach and compared it with related research on a variety of (large) real networks with encouraging results.