Dynamic Network Analysis And Behavioral Monitoring Of Terrorist Organizations

Terrorism is a major threat to national security and an important aspect of crisis management.Networked terrorist organizations are the main sources of power for the planning and implementation of terrorist activities and are the main objects of this thesis.As an important yet challenging research,Dynamic Network Analysis and Behavioral Monitoring(DNABM)extract network features from social and spatial domains,detect behavioral changes in terrorist organizations,and then provide theoretical and methodological support for effective early-warning of terrorist risks and scientific construction of counter-terrorism strategies.According to the actual background of counter-terrorism application and guided by network analysis theory,along with graph theory,time series analysis,machine learning,geostatistics,and organizational behavior principles,the framework with mutiple methods for DNABM of terrorist organizations are proposed.Taking the organizational,cross-regional,and continuously active terrorist organzaitons as the research object,the framework considers dynamic features such as correlation,periodicity,and socio-spatiality of networks and behaviors.Empirical researches of typical terrorist organizations and regions show the effectiveness and advancement of the proposed framework and methods.Based on the dynamic features of terrorist organizations and the conclusions of monitoring research,as well as the circumstances in China,corresponding counter-terrorism strategies are constructed.The main research work and contributions of this thesis are as follows.(1)The behavior-monitoring-related dynamic features of terrorist organizations are identified.Based on the mechanism of dynamic changes of terrorist organizations,the “ideal” core-periphery structure of the terrorist organization network is extracted.With the criterion of energy loss,the dynamic features of the core structure are studied.Based on tie-strength analysis,the internal rules of connection selection and cooperative reconnection in the terrorist network are studied.Based on statistical test,the abnormal structures of organizational network are compared from different aspects.Results verify the rational choice and the collaborative changes in the terrorist network dynamic process.(2)Based on the network features,an analytical framework for DNABM is proposed.On the basis of the dynamic features of terrorist organizational networks,a network-analysis-based comprehensive monitoring framework is proposed,and consists of four parts.The framework provides a complete dynamic monitoring process and specific paths and methods for correlation,periodicity,and socio-spatial dynamics monitoring.(3)Based on the correlation of the network and events,a new method for monitoring the behavior of terrorist organizations is established.Aiming at the dynamic correlationship,a monitoring method based on supervised learning and temporal pattern classification is designed.This method starts from the definition of the event-related changes of network,uses the optimized BP neural network as a classifier,and monitors the behavioral anomalies of terrorist organizations through temporal network change detection.And for the problem of unbalanced classification,a two-step monitoring method is further designed.(4)Based on the periodicity,a new method for monitoring and early-warning of terrorist organizational behavior is proposed.Aiming at the periodicity and multiple time-scale characteristics of terrorist networks,a short-term prediction and behavioral early-warning method for terrorist organizations based on wavelet transform and behavioral pattern recognition is designed.This method combines the periodicity-based qualitative inference with WNN-based quantitative analysis,and improves the confidence of the early-warning of terrorist organizations.(5)Based on the socio-spatial dynamics,a new method of monitoring and earlywarning of terrorist activities is proposed.Focusing on the socio-spatial dynamics of terrorist activities,Spatial Diffusion Networks(SDNs)of terrorist activities based on the principle of “sharing” of terrorist organizations are established from the macro perspective of socio-temporal analysis.After analyzing the diffution patterns of terrorist activities,a monitoring and early-warning method for terrorist activities based on spatial distribution,temporal test and short-term prediction is designed.(6)Based on the dynamic features of terrorist organizations and the conclusions of this thesis,the countermeasures and suggestions for improving the crisis management of counter-terrorism are proposed.Targeting the reality of counter-terrorism that China is facing,corresponding counter-terrorism response strategies are constructed.These suggestions explore the direction of counter-terrorism strategy's improvement from different aspects,such as intelligence acquisition,organizational intervention,multidomain joint monitoring,strategies adjustment,counter-terrorism cooperation,and dynamic decision-making.