Enhancing cybersecurity in communication networks using machine learning and AI: A Case Study of 5G Infrastructure Security

This study investigates the application of machine learning models for security threat detection in 5G networks, emphasizing their effectiveness in identifying malicious activities. Comparative performance analysis demonstrated that the machine learning-based approach significantly outperformed traditional signature-based detection methods, which showed a lower detection rate of 72.5%. The study also explored the trade-off between security sensitivity and operational efficiency, noting that increasing recall improves threat detection but raises false alarms, while optimizing precision reduces false positives at the risk of missing actual threats. These findings emphasize the need for a balanced security framework in 5G networks. The model was trained and validated using a dataset comprising benign and malicious network activities. The model achieved an overall accuracy of 91.5%. A confusion matrix analysis revealed that the model correctly classified 438 benign instances as non-malicious and 419 malicious instances as threats. However, 76 benign activities were misclassified as malicious (false positives), while 67 malicious activities were undetected (false negatives), highlighting a precision of 85.2% and a recall of 86.2%. The study concludes that AI-driven security models provide superior adaptability to evolving cyber threats in 5G environments. Recommendations include hybrid security approaches integrating machine learning with conventional methods, periodic retraining with updated datasets, and the development of adaptive threat management systems. These strategies will enhance detection accuracy and ensure more robust security for next-generation networks.