Random Forest-Based Intelligent Intrusion Detection and Security Monitoring for Healthcare Networks

Abstract

Healthcare cybersecurity involves protecting interconnected systems, patient information, medical devices, and digital services from evolving cyber threats. The increasing adoption of Electronic Health Records, cloud computing, telemedicine, and Internet of Medical Things devices has expanded healthcare networks and increased their exposure to attacks. This study designed and implemented an Artificial Intelligence-powered cybersecurity system using the Random Forest algorithm for intelligent threat detection, attack classification, and real-time security monitoring. A Design Science Research approach guided system development, while the CIC-IDS2017 dataset was used for experimental evaluation. The dataset was pre-processed through cleaning, duplicate removal, feature encoding, normalization, and feature selection, then divided into 80% training and 20% testing sets with five-fold cross-validation. The implemented system incorporated data acquisition, Random Forest classification, threat classification, alert generation, security logging, and dashboard monitoring. Experimental results showed 95.00% accuracy, 90.57% precision, 100.00% recall, 95.05% F1-score, 0.981 ROC-AUC, 0.912 MCC, 4.60% FPR, and 0.00% FNR, with an average detection response time of 1.9 seconds. The findings demonstrate that the proposed system provides accurate and near-real-time cyber threat detection and offers a practical approach to strengthening security monitoring and protection of healthcare networks.