ML-Guard: A Machine Learning-Based Security Device for Network Threat Defence

Cyber threats are increasing in complexity and volume, rendering traditional rule-based and signature-driven security devices insufficient.This paper introduces ML-Guard, a machine learning-based network security device designed for real-time anomaly detection, intrusion prevention, and adaptive threat response.Unlike conventional devices, ML-Guard employs a hybrid approach combining supervised and unsupervised learning for effective classification of known attacks and detection of previously unseen threats.The proposed device was tested using benchmark intrusion datasets (NSL-KDD and CICIDS2017) and demonstrated improved detection accuracy, reduced false positive rates, and robust adaptability to network changes.ML-Guard offers a scalable and intelligent architecture suitable for enterprise, IoT, and edge network environments, paving the way for next-generation autonomous security appliances.