Volume 18, No. 6, 2021
Intrusion Detection On The Unsw-Nb15 Dataset Using Feature Selection And Machine Learning Techniques
J. Vimal Rosy and Dr. S. Britto Ramesh Kumar
Abstract
Backdoor, Exploits, Shellcode, analysis, fuzzers, generic, normal, reconnaissance, DoS, and Worms assaults are among the threats detected and classified by this research study using the UNSW-NB15 dataset. To improve accuracy, the feature selection process is carried out utilising the VFS (Validated Feature Selection) algorithm, which selects just the most important features. Finally, utilising the EC (Estimated Classifier) algorithm for classification, the incursion is classified. The sort of assault is revealed during the prediction phase. Finally, the new VFS-EC model was assessed using various performance measures and compared to other existing models to demonstrate its efficacy. The findings of the study revealed that this method is highly effective in identifying and classifying attacks with greater precision.
Pages: 4784-4802
Keywords: Intrusion detection, UNSW-NB15 dataset, Feature Selection, Estimated Classifier are some of the terms used in this paper.