Abstract
Siren represents a pioneering research effort aimed at fortifyingcybersecurity through strategic integration of deception, machine learning, andproactive threat analysis. Drawing inspiration from mythical sirens, thisproject employs sophisticated methods to lure potential threats into controlledenvironments. The system features a dynamic machine learning model for realtimeanalysis and classification, ensuring continuous adaptability to emerging cyberthreats. The architectural framework includes a link monitoring proxy, apurpose-built machine learning model for dynamic link analysis, and a honeypotenriched with simulated user interactions to intensify threat engagement. Dataprotection within the honeypot is fortified with probabilistic encryption.Additionally, the incorporation of simulated user activity extends the system'scapacity to capture and learn from potential attackers even after userdisengagement. Overall, Siren introduces a paradigm shift in cybersecurity,transforming traditional defense mechanisms into proactive systems thatactively engage and learn from potential adversaries. The research strives toenhance user protection while yielding valuable insights for ongoing refinementin response to the evolving landscape of cybersecurity threats.