Machine learning techniques for intrusion detection in Smart healthcare systems: A comparative analysis

Abstract

The Internet of Things (IoT) has infiltrated nearly every aspect of life. Smart healthcare systems (SHS) are one of the most important sectors where IoT solutions are frequently deployed. With the usage of wearable devices, IoT-based smart healthcare systems have considerably improved the benefits of the healthcare industry. By combining IoT sensors with health monitoring equipment, smart healthcare minimises hospitalisation expenses and enables timely treatment for a variety of medical issues. Nonetheless, SHS is very vulnerable to a wide range of security breaches, including data leakage, data tampering, and forging. In this research, we present a comparative analysis of various machine learning-based techniques for successfully detecting intrusion in SHS. We use UNSW-NB15 network intrusion benchmark dataset to evaluate our proposed model. The evaluation results show that the AdaBoost classifier can detect network intrusion with an accuracy of 95%.