MACHINE LEARNING-BASED ANOMALY DETECTION FOR CLOUD COMPUTING SECURITY

Abstract

Cloud computing has become the backbone of modern digital infrastructures by offering scalable, flexible, and cost-effective computing resources. Despite these advantages, cloud environments are increasingly exposed to sophisticated cyberattacks, including distributed denial-of-service (DDoS), insider threats, ransomware, privilege escalation, data exfiltration, and zero-day attacks. Traditional signature-based intrusion detection systems often fail to detect previously unseen attacks because they rely on predefined attack signatures and static rule sets. Machine learning (ML)-based anomaly detection has emerged as a promising solution capable of identifying abnormal behaviors without requiring prior knowledge of attack patterns. This paper proposes a comprehensive machine learning-based anomaly detection framework designed to enhance cloud computing security through intelligent behavioral analysis and real-time attack detection. The proposed framework integrates data preprocessing, feature engineering, anomaly scoring, supervised classification, and continuous learning to improve detection performance while minimizing false alarms. Experiments are conducted using benchmark cybersecurity datasets including NSL-KDD, UNSW-NB15, and CICIDS2017. Comparative evaluation demonstrates that the proposed framework achieves an accuracy of 99.1%, precision of 98.9%, recall of 98.7%, and F1-score of 98.8%, outperforming conventional machine learning techniques such as Decision Trees, Support Vector Machines, and Random Forest classifiers. Furthermore, the proposed model significantly reduces false positive rates and detection latency, making it suitable for deployment in dynamic cloud environments. The results indicate that combining advanced feature engineering with adaptive machine learning algorithms substantially improves cloud security and supports proactive cyber threat mitigation