RID-Cloud: Spectral Recurrent Neural Network-Based Intrusion Detection in Cloud Environment

Abstract

Cloud computing (CC) is one of the most promising technologies for effectively storing data and offering internet services. There are several benefits to using this quickly evolving technology instead of more conventional defenses to shield computer-based systems from cyberattacks. Nevertheless, there are a number of drawbacks to the current approaches, including limited precision, which may affect system performance, scalability, security, and efficiency. To overcome these issues, a novel spectral recurrent neural network (RNN)-based intrusion detection in a cloud environment has been proposed for improve the security in CC. Initially, data preprocessing uses IoT-23 dataset values to reduce null or inappropriate feature values. Then the fuzzified entity is used for feature selection to analyze the features based on its threshold values. Using support index based on the behavioral rate, these threshold values select the relational features based on their maximum threshold range. Soft-max deep spectral RNN is used for identifying intrusion or non-intrusion. The intruded data is classified using recursive multi-perception neural classifier for categorized the risk level. The proposed model is superior based on accuracy, precision, recall, specificity, and F measure, according to the experimental data. The suggested tactics provide outcomes with an accuracy rate of 99.14%, which is a notable enhancement over previous studies and substantiates the effectiveness of the proposed methodology.