Enhancing Threat Detection in Healthcare Systems Through Cloud- Based Security Solutions
Pages : 180-189, DOI: https://doi.org/10.14741/ijmcr/v.12.2.4Download PDF
Growing use of digital infrastructure in the health sector has contributed enormously to healthcare but opened the systems up to an unprecedented variety of high-tech cyber-attacks. This research advances a new framework for tackling the risks through proposing a Federated Self-Adaptive Threat Detection (FedSATD) model, especially created for cloud-enabled smart health care settings. In contrast to centralized or stand-alone local traditional systems, FedSATD takes advantage of federated learning’s strengths to allow multiple healthcare institutions to jointly train a threat detection model without the exchange of patient data that is private. The decentralized architecture promotes privacy, mitigates data transfer danger, and provides regulatory compliance while enjoying disparate data distributions across institutions. FedSATD leverages deep learning approaches such as LSTM and autoencoders to scan time-series network traffic and determine cyberattack-associated anomalies. Careful preprocessing pipe is applied on each local dataset, ranging from cleaning to encoding, normalization, and sequence setup, to ensure data quality as well as prepare the model. Model performance was assessed with genuine healthcare network traces and compared to standard deep models such as DNN, GRU, and VGG-16. Experimental results indicate FedSATD works better than all the mentioned models in all of the most significant performance measures having 99.72% accuracy, 99.22% precision, 98.94% recall, and a 99.33% F1-score. Graphical plots also support FedSATD’s excellence in reducing false alarms at high detection rates. Moreover, pie charts show a high ratio of true positive detection, reflecting excellent model reliability.In summary, the FedSATD model provides a strong, scalable, and privacy-enhancing solution for bolstering cybersecurity in contemporary healthcare systems. Its adaptive learning capability from distributed environments without involving information privacy makes it a valuable contribution to the field of cloud-based healthcare security.
Keywords: Healthcare Cybersecurity, Federated Learning, Threat Detection, Cloud-Based Security, Anomaly Detection, FedSATD, Network Intrusion Detection, Cyber Threat Intelligence