Hybrid deep learning models with spotted hyena optimization for cloud computing enabled intrusion detection system

Cloud computing (CC) is Internet-assisted technology where collective resources, like platform, storage, software, and data, are presented to consumers upon requisition. CC provides scalable and dynamically adjustable resources to users over the Internet. CC presents novel vulnerabilities, with data breaches, insider threats, and unauthorized access. The standard structure of the cloud creates appealing objectives for hackers. The incorporation of solid safety procedures is vital to tackle these safety threats. One effective method is an Intrusion Detection System (IDS), which is critical in preserving cloud environments and networks. An IDS observes system activities and network traffic. Many studies have discovered the utilization of Machine Learning (ML) and Deep Learning (DL) techniques to improve IDS. They have established their capability to analyze vast numbers of data and produce precise predictions. By harnessing these models, IDSs are modified to detect prior outbreaks, develop threats, and diminish false positives. With this inspiration, this study introduces a Spotted Hyena Optimization with a Deep Learning Model for Cloud-Enabled Intrusion Detection System (SHODLM-CEIDS) technique. The presented SHODLM-CEIDS technique aims to recognize the intrusions in the cloud platform. The feature selection (FS) technique using the dung beetle optimizer (DBO) model is initially applied. Next, the SHODLM-CEIDS technique employs a Convolutional Neural Network-Bidirectional Long Short-Term Memory (CNN-BiLSTM) approach for automated and accurate intrusion detection and classification. At last, the hyperparameters of the CNN-BiLSTM method are optimally chosen by the spotted hyena optimization (SHO) model. An extensive simulation analysis is executed to inspect the improved results of the SHODLM-CEIDS model. The performance validation of the SHODLM-CEIDS technique portrayed a superior accuracy value of 99.49 % compared to other models.