This project designs and deploys a multi-disciplinary framework to model spatial, temporal and social dynamics of cyber criminals. The framework fuses theories in both computer science and criminology. Specifically, project objectives are a) Apply and validate existing theories in the realm of general criminology (in particular Akers? social learning theory and Gottfredson and Hirschi?s general theory of crime) to study cyber crimes; b) Derive novel Internet usage features as fingerprints for cyber crimes; c) Design classification algorithms (based on multi-fractal analysis and petri-net designs) to subsequently model multiple dynamics of cyber criminals by integrating theoretical and practical outcomes from the above two objectives; and d) Extensively test and validate project outcomes. The core novelty of this project is in using real Internet data from subjects (initially a cyber savvy college sample) that is collected continuously, unobtrusively, while still preserving a high degree of privacy.
Outcomes of this project will have far reaching impacts. It lays a foundation for fusing expertise in social sciences (specifically criminology) and cyber security, as a result of which existing theories in general criminology can be empirically tested for practical validity in studying cyber crimes. The identification of unique Internet fingerprints associating with cyber crimes will provide new insights into human centered aspects of cyber crimes, which is lacking today. The classification algorithms designed will provide cyber defenders with new tools to combat cyber crimes from multiple perspectives including prevention, detection, forensic investigations and prosecution.
