Jaya-Mutated Leader Algorithm Based Deep Feed Forward Neural Network for Intrusion Detection Using Cloud Computing

Abstract

Cloud computing is a compulsory choice preferred by every Information Technology (IT) organization as it is more flexible to use with wide service. However, privacy and security are major disturbances to its success because of distributed and open architecture, which is susceptible to intruders. In cloud computing, intrusion detection system (IDS) is most prominently utilized mechanism for detecting attacks. This article provides algorithm-based deep feed forward neural network (DFFNN) to detect intrusions in cloud. Here, training for DFFNN is provided by proposed Jaya-Mutated Leader Algorithm (Jaya-MLA), which is formed by integration of Jaya Optimization and Mutated Leader Algorithm (MLA). This article’s processes included for intrusion detection are data pre-processing, feature fusion, and data augmentation. Here, data pre-processing is done by Z-score normalization, feature fusion is done by deep neural network (DNN) with correlation coefficient, and data augmentation is carried out by oversampling based Synthetic Minority Oversampling Technique (SMOTE). Moreover, DFFNN-based Jaya-MLA detected intrusions and classified them into five types. Furthermore, intrusion is detected, and the model’s performance is enhanced by evaluating it with three evaluation metrics, like F1-score, recall, and precision with values of 0.948, 0.939, and 0.961.