Abstract

The exponential expansion of social media has resulted in a flood of user-generated information on previously unseen scales, making it difficult to maintain user privacy & security. This research provides a thorough analysis of hybrid machine learning (ML) models developed to identify & threats social media security risks. The review begins by exploring the landscape of social media threats, encompassing diverse categories such as cyberbullying, hate speech, misinformation, and malicious activities. Because of the potential for increased attack variety brought on by the pandemic, cyber-security will continue to be a crucial industry in the years to come. There are numerous methods to offer cyber security, including firewalls, IDS, and authentication and encryption (intrusion detection system). Subsequently, it surveys the technology on standalone ML models employed for threat detection and identifies their limitations in handling the evolving nature of threats on social media.