Detecting and Responding to Cloud Privilege Escalation Attacks Using Machine Learning

The field of cloud computing has created as an extremely productive region for complete innovative work, drawing in critical consideration from proficient organizations and organizations the same. The quick multiplication of PCs, information organizations, and other basic applications has brought about a more prominent accentuation on network safety. This exploration utilizes state of the art ML procedures to distinguish and forestall honor acceleration dangers, consequently supporting cloud security. Strong insurances are fundamental since utilizing cloud administrations builds the risk of these attacks. To work on generally speaking security, the review targets shortcomings in representative access freedoms in cloud settings utilizing ML methods like LightGBM, RF, AdaBoost, and XGBoost. This empowers ongoing discovery and relief of honor acceleration endeavors while adjusting quickly to evolving dangers. Moreover, a “soft” Voting Classifier that joins forecasts from RF, DT, and SVM supports framework execution considerably further. The blend of a simple to-utilize Flask framework with SQLite gives secure information exchange and sign-in usefulness while enhancing client testing for true arrangement. This far reaching technique not just further develops client and business information security and encourages distributed computing trust, however it likewise helps cloud specialist co-ops’ and associations’ trust in their ability to keep a protected web-based climate. Among the proposed calculations, the Voting Classifier scored the best execution accuracy of $96.45 \%$, while LightGBM accomplished $94 \%$. The leftover accuracy scores are as per the following: RF ($95 \%$), AdaBoost ($95 \%$), and XGBoost (${9 4 \%}$).