The Research And Application Of Crime Network Analysis Techniques Based On Communication Behaviors Mining

Crime network analysis (CNA) is a special kind of social network analysis (SNA), with nodes representing crime cases and/or case-relating persons. The goal of CNA is to provide effective information to law enforcement departments and intelligence agencies in crime investigation and prevention. S department proposes to establish an intelligence analysis system (TIQAS) that adopts the CNA techniques to mine communication data in order to meet the needs in organized crime investigation. This paper studies the CNA methods and techniques for communication behavior mining. We proposed the mathematical model and process model of CNA suitable for TIQAS. By analyzing the communication data of organized crime gangs, we can find out some important information regarding the organization structure, core members and action habits, and present the analysis result in visualized graphical forms.Firstly, the paper introduces the basic concepts and methods of SNA and CNA. Secondly, with the help of experiences from TIQAS project, the paper studies the CNA methods and techniques based on communication data mining, establishes the mathematical model and process model, which forms the major part of this dissertation. In particular, we adopts four times iterative query based on the theory of six degree of separation, core mining method based on shortest path algorithm, and employs the reciprocal nearest neighbor (RNN) clustering approach to find out subgroup. Thirdly, according to the overall architecture of TIQAS, the building of crime network, core mining, crime subgroup discovery and location of CNA module were designed. Fourthly, we implement the CNA module under J2EE platform and validated its functionality using the communication data from a real case. At last, a conclusion was drawn.