Criminal Detection Based on Social Network Analysis

Recent years, researchers and public security agencies are trying to solve the crime busting tasks. In this paper, we build several models to provide a probability measurement of being a criminal and show a priority list with a discriminate line separating the conspirators from the innocents. We redirect the problem of crime busting to a combined problem of community detection and information diffusion, and build well-preformed models based on the Social Network Analysis (SNA). The algorithm is tested with the data from ICM 2012. The results show that our models are adaptable to either small or large datasets.