An Effective Network Intrusion Detection Model for Coarse-to-Fine Attack Classification of Imbalanced Network Traffic

In the present day, cyber security is facing numerous attacks that are causing substantial damage to users. Recent intrusion detection systems employing advanced methods like deep learning create effective and efficient in order address these new intricate attacks. Even recent benchmark datasets trouble of prediction minority attack classes leading way missed false alarms exten- sively. Hence, biased toward coarse (majority classes) over fine (minority classes). This problem referred as Coarse Fine-Attack Classification (C-FAC). To overcome this challenge boost multi-attack classification, a novel approach has been proposed which takes advantage ensemble model phase 1 Generative Adver- sarial Networks (GAN) 2. We used classical machine classification models: Extreme Gradient Boosting (XGBoost), Decision Tress (DT), Deep Neural (DNN). GAN cast an over-sampling method enhances accu- racy The effectiveness our was evaluated using two for intrusions, namely NSL-KDD CSE-CIC- IDS2018. Based on experimental results, it found improved performance even reduced alarm rate network significantly.