Study of complex criminal networks with new statistical tools

The word “network” has become a part of our everyday language, especially after the rise of online social networks. However, human interactions are not just about socializing and having fun. Criminals also interact with each other to plan their illegal actions, especially in organized crime.

Motivated by openly available data and publicly released court documents from a law enforcement operation called “Operazione Infinito”, which was carried out in Lombardy between 2007 and 2009 to tackle the ‘Ndrangheta mafia, Professor Bocconi Daniele Durante and co -its authors developed a new class of statistical models for grouping criminals with similar patterns of connectivity, thereby shedding further light on the community structure of criminal organizations.

In fact, in most networks, not all nodes – here representing criminals – are connected to each other, and community structures usually emerge. The simplest type of community structure is characterized by dense connections within each community and more sparse connections between different communities. This corresponds to the idea that each person is more likely to connect with the people who belong to the same community.

This happens for example with classmates within a school, even though interactions between different classes are of course possible. Sometimes, interactions between different groups are even more likely than those within the same group. Consider for example two groups of animals, predators and prey, and interactions representing eating/eating.

The structure of criminal networks is much more complex. Specifically, the ‘Ndrangheta network under analysis displays a nested and hierarchical structure, with bosses at the core of the network and subsidiaries at the periphery. Both bosses and subsidiaries are then divided into subgroups that are fairly consistent with local units. Subsidiaries from each local unit interact primarily with their bosses, but then bosses from different local units communicate with each other in order to coordinate the overall operation of the organization.

“Unlike classical community detection algorithms,” explains Professor Durante, “our extended stochastic block model is able to account for these complex architectures. Furthermore, in addition to the observed connections, which here represent co-participations of criminals in encounters of the organization, we were able to incorporate additional node information derived from surveys, such as local unit and assumed role within the organization.”

“Europol recently described criminal networks as modern-day Hydras, with a complex and fluid structure. As shown in this work of ours, this type of complexity – if analyzed through appropriate statistical models – can actually be a blessing rather than a curse, and can help illuminate these dark networks”.

The study is published in The Annals of Applied Statistics.