A Novel Machine Learning Framework for Detection Iot Botnet Attacks

Abstract

In modern days, Internet of Things(IoT) plays significant role in daily activities by enabling
digital services that help in communicating with physical entities. IoT is an emerging
technology, which offers unified services and automates the operations in different domains
that range from home automation services to modern health care systems. Security is a key
concern in IoT networks, as large number of IoT devices have vulnerabilities that can be easily
exploited. Owing to the increasing usage of resource constrained IoT devices, the botnets
exploit different variations and different penetrations techniques. Many researchers have
introduced various approaches which can be deployed for detecting botnet attacks in IoT.
However, the existing approaches do not provide ample detection rate of IoT botnet attacks.
Moreover these approaches do not support in-depth analysis of wide range IoT devices network
data. To overcome the limitation of these approaches, this research work proposes a novel
machine learning framework for detection of IoT botnet attacks. The proposed framework
detects the different types of botnet attacks. The detection accuracy of the proposed ML
framework is very high as 99.34%, which significantly outperforms when compared to that of
other machine learning algorithms.