Intelligent Cybersecurity Classification Using Chaos Game Optimization with Deep Learning Model

Abstract

Cyberattack detection has become an important research domain owing to increasing number of cybercrimes in recent years. Both Machine Learning (ML) and Deep Learning (DL) classification models are useful in effective identification and classification of cyberattacks. In addition, the involvement of hyper parameters in DL models has a significantly influence upon the overall performance of the classification models. In this background, the current study develops Intelligent Cybersecurity Classification using Chaos Game Optimization with Deep Learning (ICC-CGODL) Model. The goal of the proposed ICC-CGODL model is to recognize and categorize different kinds of attacks made upon data. Besides, ICC-CGODL model primarily performs min-max normalization process to normalize the data into uniform format. In addition, Bidirectional Gated Recurrent Unit (BiGRU) model is utilized for detection and classification of cyberattacks. Moreover, CGO algorithm is also exploited to adjust the hyper parameters involved in BiGRU model which is the novelty of current work. A wide-range of simulation analysis was conducted on benchmark dataset and the results obtained confirmed the significant performance of ICC-CGODL technique than the recent approaches.