Development of an Intrusion Detection System using an Ensemble Voting Machine Learning Technique

Abstract

Intrusion Detection Systems (IDSs) are essential for identifying unauthorized access and malicious activities in network environments. The current study presents the development of an IDS utilizing a voting-based ensemble Machine Learning (ML) approach. Utilizing the advantages of individual ML models, the voting classifier is a well-known ML model that may enhance overall prediction performance. This study provides a unique classification method that combines the benefits of the Naive Bayes (NB), K-Nearest Neighbors (KNN), and Adaptive Boosting (AdaBoost) algorithms into a voting ensemble approach. This ensemble voting classifier greatly improves network IDS accuracy. The experiments were conducted using the KDD99 dataset. The findings reveal that the voting ensemble technique outperforms individual classifiers, achieving a higher accuracy of 99.79%.