A Topic Clustering Method to Identify Online Threats against Soft Targets
Marco San Biagio, Marco Cipolla, Ernesto La Mattina, Vito Morreale · 2023
In today's rapidly changing digital world, and with the increasing prevalence of AI-based technologies, it is crucial to protect vulnerable and easily targeted locations from potential threats. These vulnerable locations, known as “soft targets,” include various public spaces and institutions that are at high risk for security breaches. To address this issue, this study aims to introduce an AI -driven application that aids intelligence analysts and investigators in identifying security threats on social media platforms in a timely manner. The application utilizes a topic modelling algorithm trained on a comprehensive dataset related to terrorism activities. It identifies and suggests trending topics to enhance the monitoring process, thereby improving the quality of information collected from social media platforms. This approach empowers security professionals and decision-makers, even those with limited expertise in the field, to gain valuable insights and strengthen the protection of public spaces and soft targets. The results obtained from testing the solution on a newly generated dataset demonstrate its efficiency and effectiveness, reinforcing its significance in safeguarding vulnerable locations from potential security threats.