AI-Based Methods to Detect and Counter Cyber Threats in Cloud Environments to Strengthen Cloud Security
Diwakar Chaudhary, Sanjeev Verma, Vijay Mohan Shrimal, Ravikiran Madala, Rashi Baliyan, Sathish Muthu
Abstract
Traditional cloud security solutions often employ outdated techniques and find it difficult to stay up to date with the ever-evolving cyber threat scenario. Complex attacks may be made against cloud systems since static signatures and rule-based methods have their limits. To address these problems, the study proposes a novel AI-based approach to enhance cyber threat detection and mitigation in cloud settings. By use of sophisticated anomaly detection and machine learning (ML) methods, the proposed system continuously analyzes large amounts of data to dynamically identify and eliminate new threats. Some of the key components include robust preprocessing, a variety of ML model selections, repetitive model training, real-time threat detection, and automated response mechanisms. Compared to the current system, the results demonstrate notable increases in detection accuracy, false positive and false negative rates, and area under the curve (AUC) values across many test datasets. Specifically, the proposed system reduces the false negative and false positive rates to 2.0% and 1.2%, respectively, and achieves 98.5% detection accuracy as compared to 89.7% for the existing system. Moreover, on several test datasets, the proposed system routinely produces higher AUC values than the existing system, proving its efficiency in improving the accuracy of cyber threat identification in cloud systems.