Data dependent similarity-based machine learning approach to criminal intelligence problems
PhD
Data dependent similarity-based machine learning approach to criminal intelligence problems
Outline
Criminal Intelligence is information compiled, analysed, and/or disseminated in an effort to anticipate, prevent, or monitor criminal activity. Rapid advances in data acquisition have created new challenges and opportunities for practitioners, including profiling and predictive policing, forecasting crime risk and crime rates, and legal and adversarial decisions. The underlying machine learning tasks for this problem set are difficult because of the high dimensionality, but more importantly the ill-structuredness and of the domain, requiring a more sophisticated and adaptable notion of similarity. This project aims to create a new similarity-based machine learning approach to solve this problem by making use of existing state-of-the-art data dependent similarity measures.
Supervisory Team
Principal Supervisor: Dr Giles Oatley
Co-supervisors:
Dr Suryani Lim