An Improved Design for a Cloud Intrusion Detection System Using Hybrid Features Selection Approach with ML Classifier

The focus of cloud computing nowadays has been reshaping the digital epoch, in which clients now face serious concerns about security and privacy their data hosted cloud, as well increasingly sophisticated frequent cyberattacks. Therefore, it become imperative for both individuals organizations to implement a robust intrusion detection system (IDS) capable monitoring packets network, distinguishing between benign malicious behavior, detecting type attacks. IDS based on ML are efficient precise spotting network threats. Yet, large dimensional sizes, performance these systems decreases. Thus, is critical building suitable feature selection approach that selects necessary features without having an impact classification process or causing information loss. Furthermore, training models unbalanced datasets show rising false positive rate (FPR) lowering (DR). In this paper, we present improved designed by incorporating synthetic minority over-sampling technique (SMOTE) address imbalanced issue, selection, propose use hybrid includes three techniques: gain (IG), chi-square (CS), particle swarm optimization (PSO). Finally, random forest (RF) model utilized classifying various types suggested verified UNSW-NB15 Kyoto datasets, achieving accuracies over 98% 99% multi-class scenario, respectively. It was noticed with fewer informative would operate more effectively. simulation results significantly outperform other methodologies proposed related work terms different evaluation metrics.