Machine Learning-Based Intrusion Detection for Cloud-Connected Electronic Manufacturing Systems
Keywords:
Intrusion Detection; Machine Learning; Cloud-Connected Manufacturing; Industrial Cybersecurity; Network Anomaly Detection.Abstract
This study proposes a machine learning-based intrusion detection framework for cloud-connected electronic manufacturing systems, focusing on improving cybersecurity, network reliability, and operational resilience. The proposed methodology integrates network traffic monitoring, industrial communication logs, and cloud access records to identify malicious activities affecting interconnected manufacturing environments. Advanced data preprocessing and feature selection techniques are employed to extract relevant traffic characteristics, including packet frequency, connection duration, protocol behavior, and abnormal access patterns. Supervised learning algorithms, including Random Forest, Support Vector Machines, and Gradient Boosting, are investigated for detecting and classifying network intrusions. An adaptive anomaly detection mechanism is incorporated to identify previously unseen attack patterns and reduce dependence on predefined threat signatures. Performance evaluation considers detection accuracy, precision, recall, F1-score, false positive rate, and computational latency. The proposed framework supports early cyberattack identification, secure cloud integration, and reliable industrial operations through intelligent threat monitoring and automated security decision support.