Threat Eye: Behavior Analytics for Cloud Security using ML

Cloud computing has developed for its storage and accessibility of data, providing flexibility, scalability, and affordability. But its extensive use has also resulted in cyberattacks and unauthorized access. The traditional intrusion detection mechanisms become ineffective for different attack patterns, insider attacks and zero-day exploits. This paper provides an efficient way to address this issue by providing a mechanism to identify the threat based on the behaviour of the user. Unlike traditional approaches this paper provides a system that monitors the user behaviour patterns considering the features such as login history, authentication attempts, session duration, file access activities, password change frequency and IP location modifications. By considering these features the attack can be detected, and the data access is restricted to prevent from unauthorized access. This paper provides handling of attack with a combination of machine learning algorithms including SVM, Decision tree and XBoost, these help in distinguishing between the attacker and the authorized user. This helps in identifying complex attack behaviours and detect insider threats. Additionally, RSA encryption is integrated to secure data transmissions, preventing unauthorized access. By combining behavioural analysis, attacks prediction, and encryption, this provides a robust cybersecurity framework.