Machine Learning-Based Detection and Mitigation of Privilege Escalation Attacks in Cloud Environments

There are serious security dangers in cloud systems due to the quick rise in cyberthreats, especially privilege escalation assaults. Because cloud computing is centralized, it is susceptible to insider attacks, in which authorized people abuse their authority to obtain unapproved access. Insiders have authorized access, unlike external attackers, which makes it more difficult to identify their nefarious activity. This paper suggests a machine learning-based approach that combines several cutting-edge techniques to identify and prevent privilege escalation attacks. A methodical technique is created to examine user behavior and spot unusual activity that might point to security lapses. To improve model performance, the project uses hyperparameter tuning, SMOTE-based data balancing, and feature engineering. Using a stacking classifier that combines Random Forest, XGBoost, LightGBM, CatBoost, and Gradient Boosting in order to use ensemble learning.