Politecnico di Torino - Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY

+39 011 090 6100 info@tech-share.it

METHOD FOR PREDICTIVE ANALYSIS OF THE SOCIAL STRUCTURE OF CRIMINAL ORGANIZATIONS

Criminal graphCriminal organizationsMultilayer graphMultilayer networksPredictive analysis

Introduction

The invention refers to a method for the predictive analysis of the social structure of criminal organizations. Specifically, the invention is represented by a software platform for the multidimensional analysis of the structure of criminal organizations (for example mafia gang or n’drangheta clan). The proposed system aims to identify the most dangerous individuals operating within criminal organizations and to predict future collaborations between individuals aimed at carrying out criminal actions.

Technical features

The method of the invention is represented by the following steps:

– Identification of a plurality of levels of relationships between subjects belonging to or associated with a criminal organization;

– Construction of a plurality of graphs able to describe relationships between the subjects described above, where each node represents a subject and where each graph represents a level of relationship between subjects and the graphs are related to forming a multilayer structure of graphs;

– Construction of a plurality of links between the subjects, where said links connect subjects on the same graph or on a different graph belonging to the same multilayer structure of graphs;

– Execution of a link prediction algorithm in order to calculate the probability of the establishment of a further link between subjects in a future moment of time.

Possible Applications

  • Criminology;
  • Computer technology;
  • Investigation activities;
  • Intelligence agencies and Police Forces;
  • Social sciences.

Advantages

  • Exceeds the limits deriving from telephone interceptions only;
  • He examines various investigative sources, using a complex and flexible mathematical structure, that is the multilayer graph;
  • Limits the intervention of the human operator and relies on statistical algorithms.