A Comparative Study of Machine Learning Models for Intrusion Detection in 5G Traffic
Ayush Ayush, Sandeep Kumar Singh · 2024
The expanding use of 5G networks necessitates robust security measures due to their intricate architecture and enlarged attack landscape. Intrusion detection systems (IDS) are crucial for shielding these networks by pinpointing malicious traffic. This study explores the efficacy of machine learning models for intrusion detection in 5G traffic. This research employs three established machine learning algorithms: Random Forest (RF), XGBoost, K-Nearest Neighbors (KNN) and Histogram-based Gradient Boosting(HistGradient Boosting). We conduct a comparative analysis on both datasets, assessing their performance metrics like accuracy, precision, recall, and F1-score. Feature selection techniques are implemented to enhance model performance and address potential dimensionality challenges. The research tackles the class imbalance issue commonly found in intrusion detection datasets, ensuring a fair evaluation of the model’s ability to detect both normal and malicious traffic. The findings offer valuable insights into the suitability of these machine learning models for 5G network intrusion detection. We compare the strengths and weaknesses of each model across the datasets, highlighting their effectiveness in identifying diverse attack types. This research contributes to the ongoing development of reliable and efficient IDS solutions specifically designed for the security needs of 5G networks. The highest achieved accuracy is 99.97% for binary classification of "Attack type" to detect intrusion.