Hybrid Intrusion Detection for Malware Threats in Cloud Computing using Principal Component Analysis and Harris Hawks Optimization

Abstract

The increasing threats from malware are challenging defensive mechanisms in the continuously evolving world of cloud computing. Malicious malware now poses one of the biggest challenges towards cloud data security. To overcome this, a novel Cloud Malware Intrusion Detection System has been proposed in this paper using a hybrid approach incorporating Harris Hawks Optimization along with Multilayer Perceptron and Principal Component Analysis. This hybrid methodology integrates the capability of PCA for dimensionality reduction so that the extraction of significant features is allowed to occur efficiently, while HHO optimizes hyperparameters for MLP so that accuracy and performance are improved. By leveraging deep learning in order to detect subtle abnormalities indicative of malware threats, the core of a vigilant data monitoring system was found within the cloud architecture as MLP. It has been discovered that the hybrid PCA-HHO-MLP model is an incredibly effective and flexible security mechanism against the constantly changing malware field. According to experimental results, the suggested approach may successfully protect cloud resources while providing exceptional detection performance with accuracy: 98.45% , precision: 98.55% , recall: 99.88% , and F1 Score: 99.21% .