Explainable AI-Driven Adaptive Healthcare 5.0 Security Framework for Intelligent Intrusion Detection and Secure IoMT Communication

Resumen

The Internet of Medical Things (IoMT) has significantly expanded the cyber-attack surface area of Healthcare 5.0 environments, with the widespread use of Internet of Things (IoT) devices, cloud systems, remote monitoring platforms, and intelligent clinical systems. This research introduces an integrated, explainable and adaptive security framework for intelligent intrusion detection and secure IoMT communication scenario. The framework includes four components: Isolation Forest, a unsupervised anomaly detection algorithm; Random Forest, a supervised intrusion and access classification algorithm; SHAP, a feature-level explainability algorithm; and Adaptive Risk Optimization, an algorithm for making security decisions in context; and AES–Fernet is an encryption algorithm for protected communication. The suggested workflow does not consider these functions as stand-alone security measures, but connects anomaly evidence with user behaviour, device trust, risk assessment, authorization, explanation and secure communication. The accuracy, precision, recall, and F1 score derived from the independent test dataset were determined to be 99.84%, 99.38%, 99.69%, and 99.53%, respectively. The results show that the integrated framework is able to achieve high and stable performance in intrusion detection and solve the transparency and adaptive security issues. The main contribution is architectural and methodological: existing algorithms are orchestrated in a sequential pipeline in which the detection process is followed by risk assessment, the risk process is followed by security decision-making, and SHAP supplies interpretable evidence to support security decision-making. The framework thus provides a feasible basis for a transparent, context-aware, and secure cybersecurity in the context of Healthcare 5.0, and its feasibility in the deployment process is subject to the validation of real-world hospital and diverse IoMT data.