Abstract
The Internet of Things (IoT) age is quickly evolving, with millions of
devices and many more intelligent systems, like healthcare. Attackers mostly aim for
these IoT devices. These devices are infected with malware, which turns them into bots
that are used by attackers to disrupt networks as well as steal important data. To
address this issue, efficient machine learning combined with appropriate feature
engineering is proposed to detect and protect the network against vulnerabilities. The
proposed model will detect Distributed Denial of Service (DDOS)-based botnet attacks
in the smart healthcare system. Hacktivists frequently use DDoS assaults to overwhelm
networks and make them unusable. For healthcare providers who depend on network
connections to enable efficient patient data access, this can be a serious problem. DDoS
attacks are motivated by a social, political, ideological, or economic motive tied to a
scenario that enrages cyber threat actors. Two modern Machine Learning (ML)
methods, including (i) Support Vector Machine (SVM) and (ii) Light Gradient
Boosting Machine (Light GBM), are used to validate the data set. From the extensive
experimental analysis, feature-based algorithms are superior to other competing models
in that they (i) have the highest detection rate with high accuracy, and (ii) have less
computational complexity with minimal training and test time.