A DEEP LEARNING BASED INTRUSION DETECTION SYSTEM FOR HEALTHCARE APPLICATIONS

Abstract

The innumerable increase in user information and network traffic in healthcare applications has made it complex, especially for network-intrusion-detection systems (NIDS) to be familiar with and perform well. Therefore, Intrusion Systems are considered pivotal in e-healthcare systems since the medical details of patients should be confidential, precise, and highly secure. Any variations in the original patient record can lead to massive changes or lead to faults in the treatment as well as diagnosis of diseases like brain stroke. Further, most of the existing studies are focused on intrusion-based systems that are trained with outdated records and intrusion-detection repositories that can generate false positives at a higher rate and need to retain the technique from scratch to handle the complexities that persist in new attacks. In the current research, a novel hybrid method has been integrated with the Gated Recurrent Unit (GRU) along with the Ant Colony Optimization (ACO) algorithm for intrusion detection in healthcare applications. The proposed approach outperforms the existing models by achieving a phenomenal accuracy of 99%.