Botnets Attack Detection Using Bio‐Inspired Deep Learning Techniques in Internet of Medical Things (IoMT)

Abstract

Listen

Translate article

ABSTRACTAccording to the experts, the Internet of Medical Things (IoMT) is the next big thing and a revolutionary technology that is fast and provides more accurate diagnoses to deliver efficient healthcare services with reduced costs. The IoMT network mostly consists of low computational power devices, which makes them an easy target for serious security and privacy threats, one of which is botnet attacks. Botnet attacks pose a severe threat to IoMT, and detecting them in a timely and accurate manner is crucial for maintaining the confidentiality and integrity of sensitive medical data. A botnet in IoMT leads to attacks on confidentiality, authenticity, integrity, and availability of data and resources. The existing approaches have failed to accurately identify and detect botnet attack traffic in the IoMT environment. To accurately identify and detect botnet attacks in the IoMT environment, we proposed deep learning techniques. This is a novel botnet attack detection system for IoMT that utilizes feed‐forward neural networks (FFNNs) and convolutional neural networks (CNNs). First, we train the system using FFNN to identify and extract relevant features from the network traffic generated by IoMT devices. We then retrain the system using CNN to enhance its accuracy and performance in detecting botnet attacks. In this regard, we achieved a high accuracy of 99.94%, which is a notable achievement. To assess the overall performance, this study has incorporated various important metrics, such as accuracy (99.8%), F1 score (1.00%), specificity (0.998), sensitivity (0.997), precision (1.00), and ROC AUC (0.998). For a secure and reliable IoMT, detection is insufficient; we must also take steps to prevent IoMT attacks.