Abstract

Cybersecurity in healthcare systems is crucial due to the sensitive nature of patient data and the increasing risk of cyberattacks. This paper proposes a novel approach to enhancing cybersecurity defenses by utilizing stacked artificial neural networks (ANNs) for the detection of man-in-the-middle (MITM) attacks, such as spoofing and data injection, targeting healthcare data transmission. The dataset used includes 35 network flow metrics and 8 biometric features, which represent the critical aspects of network traffic and patient health data. The methodology involves training two base-level ANN models: one for detecting anomalies in network flows and the other for identifying irregularities in biometric data. These models' outputs are combined in a meta-model to improve classification accuracy and provide a robust defense mechanism. The proposed method demonstrates enhanced attack detection capabilities by leveraging both network and health-related data, offering a comprehensive solution to safeguard healthcare information systems.