The escalating volume of data exchange across networked devices has propelled the need for robust intrusion detection systems (IDS) capable of swiftly identifying and mitigating emerging threats. Leveraging machine learning algorithms, this study presents a Progressive Intrusion Detection System (PIDS) designed to efficiently analyze vast datasets, detect anomalous behaviors, and promptly respond to potential network intrusions. The system evaluates four distinct attack types—Support Vector Machine (SVM), Naive Bayes, Logistic Regression, and an ensemble model comprising XGBoost and Decision Trees—based on their predictive performance and adaptability to class labels. Performance evaluation metrics including F1-Measure, Accuracy, Precision, and Recall are employed to gauge the efficacy of each model. Results indicate that the ensemble model, AdaBoost with Logistic Regression, exhibits superior performance compared to alternative approaches investigated in this study. Comparative analysis with existing research demonstrates the efficacy of the proposed IDS in outperforming current state-of-the-art solutions. Finally, this paper discusses pertinent challenges and outlines future research directions for advancing intrusion detection capabilities.
Keywords: PIDS, AdaBoost, IDS, XGBoost.