Paper Contents
Abstract
The rapid expansion of cloud computing and smart device usage has significantly heightened cybersecurity challenges, particularly related to privilege escalation attacks. Centralized cloud services create critical vulnerabilities, making systems prone to both accidental and malicious data breaches. Among potential threats, malicious insiders pose the greatest risk due to their legitimate access, which provides numerous opportunities to inflict substantial damage. An advanced machine learning-based system is proposed to detect and mitigate privilege escalation attacks within cloud environments. This system leverages ensemble learning techniques, including Random Forest, AdaBoost, XG Boost, and Light GBM, to systematically identify and classify anomalous activities that may signal insider threats. A customized dataset derived from the CERT dataset is employed to evaluate the effectiveness of these models, with Light GBM achieving the highest accuracy. Despite the strong performance of Light GBM, it is essential to incorporate multiple machine learning algorithms to ensure robust detection across various insider attack scenarios. Insights from a systematic review of cloud security threats and machine learning methodologies further underscore the necessity of hybrid approaches to improve detection and mitigation strategies. Integrating past decision data with current machine learning outcomes, applied through supervised learning models across different datasets, offers a promising path toward enhancing the resilience of cloud security frameworks.
Copyright
Copyright © 2024 Jami Pradeep. This is an open access article distributed under the Creative Commons Attribution License.
