Publication

Abstract : The rapid evolution of cyber-attacks has heightened the need for advanced Intrusion Detection Systems (IDS) capable of detecting a wide range of network-based threats. This paper investigates the effectiveness of three machine learning algorithms—Decision Tree, Random Forest, and Support Vector Machine (SVM)—for identifying network intrusions using the UNSW-NB15 dataset. The dataset, rich in both normal and attack traffic instances, is used to train and test each model. The models are evaluated based on performance metrics including accuracy, precision, recall, and F1-score. Among the three algorithms, Random Forest demonstrates superior performance, showing the highest accuracy, precision, and recall. This paper contributes to the IDS research by providing a comparative analysis of these machine learning models on a modern and diverse dataset, highlighting the strengths and weaknesses of each approach for practical intrusion detection.