A user interface for fraud detection systems

The interface will be developed in Wolfram Mathematica so that the user will only have to download the Wolfram CDF Player to run the implementation on a web browser (without the need of a Mathematica license or any other associated cost). To our knowledge the proposed user interface would be novel in that it focuses on quantifying the impact of empirical community structure on detection performance, according to different measures. The overall objective is align with other efforts of our research group of designing algorithms that exploit structural information (e.g., the design of spectrum-based detection algorithms). The students participating in the development of the user interface will become familiar with the network theory and the theoretical foundations underlying the algorithms. Hopefully this will allow them to develop a keen intuition that will be important for future research on fraud detection systems.