Cyber-Attack Detection for Cloud-Based Intrusion Detection Systems

Abstract

In this research, we delve into an exhaustive ex- amination of the usage of machine learning (ML) models for the identification of cyber threats within Cloud-Based Intrusion Detection Systems (IDS). With the escalating dependence on cloud services across various industries, the urgency to develop effective and resilient IDS to mitigate burgeoning cyber risks is paramount. Our study makes a significant contribution to this imperative field by analyzing the efficacy of diverse ML models in detecting cyber-attacks within a cloud context. We scrutinized an array of ML models, namely Decision Trees (DT), Random Forest (RF), XGBoost, and Support Vector Machines (SVM), utilizing key performance parameters such as accuracy, recall, precision, F1-s, and confusion matrix to understand their practical application in real-world IDS scenarios. XGBoost stood out as the most proficient model, showcasing not only an impressive accuracy but also a balanced performance in terms of precision and recall. This highlights the considerable potential of ensemble and gradient boosting techniques in optimizing cloud- based IDS detection capabilities. Our findings underscore the significant role of machine learning in fostering more dependable, robust, and efficient IDS in the cloud, thus significantly aiding in securing our digital ecosystems.