Dynamic surveillance: A case study with Enron email data set

Surveillance is a critical measure to break anonymity. While surveillance with unlimited resources is often assumed as a means, against which, to design stronger anonymity algorithms, this paper addresses the general impact of limited resource on surveillance efficiency. The general impact of limited resource on identifying a hidden group is experimentally studied; the task of identification is only done by following communications between suspects, i.e., the information of whos talking to whom. The surveillance uses simple but intuitive algorithms to return more intelligence with limited resource. The surveillance subject used in this work is the publicly available Enron email data set, an actual trace of human interaction. The initial expectation was that, even with limited resource, intuitive surveillance algorithms would return the higher intelligence than a random approach by exploiting the general properties of power law-style communication map. To the contrary, the impact of limited resource was found large to the extent that intuitive algorithms do not return significantly higher intelligence than a random approach.