AI-Enabled IDS for Identifying Malicious Traffics in Cloud Systems

The widespread usage of the Internet has increased network security issues. Some of the serious threats are Brute force, Heartbleed, Botnet, Web attacks, DoS, DDoS, and Infiltration of the network. With the increased usage of online services and the development of new technologies like IoT, Cloud, Edge, and Mobile Computing, the detection of these threats is important and becoming paramount. This study presents an Intrusion Detection System (IDS) framework by proposing various hybrid models for the detection of malicious traffic. The models are developed using feature selection techniques, namely Recursive Feature Elimination, Chi-Square, Principal Component Analysis, and five Machine Learning (ML) techniques. In this experiment, the standard CSE-CIC-IDS 2018 dataset is used to examine the performance evaluation of the proposed approaches. In this paper, according to the comparison of the proposed hybrid approaches, Recursive Feature Elimination with Random Forest performs better than the other approaches with an accuracy of 99.23%. The proposed approach provides better accuracy as compared to earlier reported similar approaches in the literature.