Machine Learning Enhanced Healthcare Cybersecurity Framework for Secure Medical Information Management

Abstract

The increasing digitalization of healthcare, protecting sensitive patient information and healthcare systems has become a critical challenge. The confidential nature of medical data and the growing number of cyber threats highlight the need for advanced cybersecurity solutions. This research proposes a Healthcare Cybersecurity Framework using machine learning and deep learning methods to handle Medical Information securely and detect cyber-attacks effectively. The proposed framework makes use of the WUSTL-EHMS-2020 dataset and preprocessing techniques such as data cleaning, normalization, label encoding, and SMOTE-based sampling. The accuracy, precision, recall and F1-score are employed for the model development and evaluation in DNN and XGBoost. In terms of accuracy, precision, recall, and F1-score, the experimental findings show that the suggested DNN model reached 99.98%, while the XGBoost model reached 99.96%, PRE 99.91%, REC 99.95%, and F1-score 99.97%. The findings show that the suggested architecture successfully increases the ACC of threat detection and safeguards the medical data management process in terms of security.