Detecting Hidden Hierarchy In Terrorist Networks Based On SNA

Since the911in USA, the study of terrorism has become a hot issue togovernments. Analyzing the structure of terrorism correctly will help governments totake counter-terrorism actions effectively. Information collection system’s constructioncan give governments more information about terrorism than before. It is not sufficientto deal with the growing data with traditional manual analysis, so analysis of thecollected data by computer automation is an urgent need for counter-terrorism actions.In the paper, the social network analysis technologies are taken into the study ofterrorist organizations. Centrality measurement, the traditional social network analysis,is a basic approach for analyzing a mass of data, so this paper will first introduces someclassical centrality measurements. With our research we can know the defect of thesecentrality measurements is that these measurements only considers the associationamong nodes and ignores the distinction of different strength among ties. To improvecentrality measurement’s performance, this paper proposed the dependencemeasurement which is a new centrality measurement considered nodes’ tie-strength.This measurement first need calculate the dependence among nodes, and distinguish thedifference of the tie-strength among nodes, so it can represent accurately the conditionof central node in the network. The experimental result shows that our newmeasurement’s performance is better than before measurements.It is first step to find the core member of terrorist organization by the calculation ofnode centricity. Then, this paper will design the algorithm of detecting hierarchicalstructure in terrorist organization. There is a growing amount of literature on modelingterrorist networks as graphs. However modeling these networks as graphs ignores animportant aspect of their structure, their hierarchy, and the fact that they are composedof leaders and followers. This algorithm uses tree model to describe terroristorganization, rather than traditional graph model, so it can easily extract the hierarchy interrorist organizations.