Detection of Attacks in Smart Healthcare deploying Machine Learning Algorithms*

The Internet of Things (IoT) is a sort of network that uses a set of protocols and data-sensing tools to link anything to the Internet. This type of network allows for data exchange as well as smart recognition, surveillance, deployment, tracking, and maintenance. The healthcare business is rapidly digitizing, and the IoT is a key factor in this. It has already had an impact on medical networks and will continue to do so soon. This paper highlights privacy concerns as well as security attacks, which have rapidly increased in recent years in the healthcare industry. An architecture is proposed for detecting IoT attacks in smart health-care systems using machine learning (ML) techniques. This study used the UNSW-NB15 dataset to simultaneously classify Random Forest (RF), Naive Bayes (NB) and K-Nearest Neighbour (KNN) classifiers for detecting the attacks (Generic, Fuzzers, Exploits, Analysis, Denial of Service (DoS), Reconnaissance, Backdoor, Worms and Shellcode) in the healthcare industry. In addition to developing the essential smart and effective online system, this research increased knowledge of privacy and security in the smart healthcare sector. Furthermore, performance measures like accuracy, precision, recall, and F1-score have been calculated using the existing ML techniques based on the mentioned UNSW-NB15 dataset. The results indicate that when comparing the mentioned ML techniques, the KNN model has the highest accuracy value, or 90%, while RF and NB provide 89% and 75% accuracy values, respectively.