Intrusion detection system using combination of deep residual fuzzy network and white shark-dwarf mongoose optimization

Abstract

Intrusion Detection Systems (IDS) are utilized to identify malicious attacks and restore a secure state for the network, in cloud environment. Cloud assisted IDS having hybridized components from Machine Learning models has demonstrated better performance on parameters like Accuracy, DR and FPR. Machine Learning techniques facilitate accurate categorization and attack prediction with different IDS related databases. Some issues which need to be addressed are the selection or fusion of the sub set of features to be selected for analysis and detection of attacks considering high data sizes available and design of appropriate classification model. Using chosen features, computational Intelligence techniques provide better efficacy on a reiterative basis. The main issue to be addressed is the choice of a suitable scheme that provides better outcomes. This paper presents a Hybridized Meta-heuristic model that unified Deep Quantum Neural Network using Angular Separation Distance for feature fusion, White Shark- Dwarf Mongoose Optimization based on Fuzzy Classifier for parameter optimization, and Deep Residual Network classification model, to design effectual IDS in cloud platform. The proposed model performed with accuracy of 90.6%, F1-score of 90.2%, precision of 90.5% and recall of 90.5%.