SMMO-CoFS: Synthetic Multi-minority Oversampling with Collaborative Feature Selection for Network Intrusion Detection System

Abstract Researchers publish various studies to improve the performance of network intrusion detection systems. However, there is still a high false alarm rate and missing intrusions due class imbalance in multi-class dataset. This imbalanced distribution classes results low accuracy for minority classes. paper proposes Synthetic Multi-minority Oversampling (SMMO) framework by integrating with collaborative feature selection (CoFS) approach Our aims increase extreme (i.e., user-to-root remote-to-local attacks) improving dataset’s selecting relevant features. In our framework, SMMO generates synthetic data iteratively over-samples multi-minority And collaboration correlation-based an evolutionary algorithm selects essential We evaluate random forest, J48, BayesNet, AdaBoostM1. NSL-KDD dataset, experimental show that proposed significantly improves compared other approaches.