Enhancing Security with Machine Learning-Based Intrusion Detection in 5G Networks

---

Abstract

Tracking network traffic and searching for anomalies is the function of intrusion detection system (IDS) software. Unusual or abnormal network changes could indicate fraud at any stage, from the initial attempt to the full-blown invasion. Data sharing needs to be secured because it mainly relies on the internet.Data encryption and authentication alone are insufficient for internet security, and firewalls cannot detect fragmented fraudulent packets. Furthermore, attackers frequently modify their plans, tools, techniques, and strategies, which can have disastrous consequences such as decreased productivity, financial loss, data loss, and so on. This paper proposes an intrusion detection system (IDS) to classify attacks on 5G networks. The classification is carried out using well-known machine learning (ML) classifiers. The proposed IDS is evaluated using standard performance metrics, which provides valuable insights into the IDS's strengths and weaknesses. Random Forest outperforms all other classifiers in terms of accuracy, with a rate of 99.8%. Random Forest has the highest accuracy for classifying attacks, at 99.89%. The high accuracy of Random Forest undergoes a potential component in developing a robust IDS for identifying and mitigating cyber threats in the 5G networks.

---

Keywords

Intrusion detection system; machine learning; confidentiality; integrity; availability