Enhanced Cloud Intrusion Detection Using Support Vector Machines

The broad adoption of cloud computing (CC) as a new technology offering on-demand networked access to shared configurable computer resources has resulted in an increased need for data storage and service management. But at the same time, CC also opens up further vulnerabilities by throwing issues such as insider threats, data breaches and unauthorized access into the mix. The study suggests this mechanism as a cloud-based intrusion detection technique made up of optimization and machine learning algorithms that mitigate these issues. In the beginning, an input noiseless training data undergoes a Gaussian filter to improve the quality of data and remove any background noise. In this step, Random Forest Recursive feature elimination is used to select significant low-dimensional features with the help of correlation-based feature selector. Implemented on the Python platform using the CICIDS-2019 dataset, the proposed method's accuracy is validated and evaluated against various criteria, demonstrating superior performance compared to recent related studies and effectively enhancing intrusion detection in cloud computing.