A real-time model for analyzing online social networks and predicting terrorist groups behavior

The world continues to face a growing threat from radicalized violent extremists. Therefore, law enforcement agencies need to be able to identify and assess the risks of large numbers of vulnerable individuals to prevent violent extremism in the future. A system-based approach is presented in this article to achieve the desired goal using machine learning models and Graph pattern matching techniques. this objective technology is described as a 'radicalization detection system,' which assists intelligence analysts in identifying extremists. Additionally, the goal is to develop a computational tool that can discover and track radicalization indicators across large databases as part of a nested effort to demonstrate technological opportunities for a radicalization detection system. As part of this technical section of the tool, a dynamic graph pattern matching approach is used to identify individuals or groups at risk of violent extremism.

Keywords:

Graph Pattern-Matching , sentiment analysis , Generative Adversarial Networks (GAN) , Outlier Detection , Parallel processing