Revolutionizing Intrusion Detection: An Incremental Majority Voting Strategy with Machine Learning

With the rapid growth of digitalization and the increasing volume of data, the cybersecurity threat landscape is expanding at an alarming rate. Intrusion Detection Systems (IDS) have become crucial in conjunction with firewalls to safeguard networks from malicious activities. In this work, four well-known cybersecurity datasets—CIC IDS 2017, NSL KDD, KDD Cup, and CIC IDS 2018—are employed to evaluate the effectiveness of various techniques for intrusion detection. Feature selection is performed using Mutual Information to enhance the relevance of selected features. Data sampling techniques are also explored, including Original Data, Random Under Sampling, Random Over Sampling, and a combination of both under and over-sampling to address data imbalance. To further improve the detection performance, a refined approach utilizing a Stacking Classifier combining Random Forest (RF) and Decision Tree (DT) with a Bagging Classifier is implemented. The results show that this approach achieves high performance across all datasets and sampling techniques, demonstrating its effectiveness in accurately detecting network intrusions in dynamic cybersecurity environments. “Index Terms – Incremental learning, network intrusion detection, machine learning, majority voting classifier, random sampling, Stacking classifier, Cyber Security”.