Machine Learning Methods for Detecting Internet-of-Things (IoT) Malware

Machine Learning Methods for Detecting Internet-of-Things (IoT) Malware

DOI: 10.4018/IJCINI.286768
Article PDF Download
Open access articles are freely available for download

Abstract

This study aims to analyze the performance of machine learning models for detecting Internet of Things malware utilizing a recent IoT dataset. Experiments on the IoT dataset were conducted with nine well-known machine learning techniques, consisting of Logistic Regression (LR), Naive Bayes (NB), Decision Tree (DT), k-Nearest Neighbors (KNN), Support Vector Machines (SVM), Neural Networks (NN), Random Forest (RF), Bagging (BG), and Stacking (ST). The results show that the proposed model attains 100% accuracy in detecting IoT malware for DT, SVM, RF, BG; about 99.9% percent for LR, NB, KNN, NN; and only 28.16% for ST classifier. This study also shows higher performance than other proposed machine learning models evaluated on the same dataset. Therefore, the results of this study can help both the researchers and application developers in designing and building intelligent malware detection systems for IoT devices.
Article Preview
Top

Introduction

Internet of Things (IoT) is a collection of interconnected devices embedded with light processors and network cards capable of being managed over the Internet (Moustafa, Turnbull, & Choo, 2018a). It represents a network of physical objects (things) embedded with sensors, software, and other technologies to exchange data with other devices and systems (Rouse, 2019). Internet of Things comprises typical network elements (workstations, laptops, and routers), sensors, actuators, smart devices, and radio frequency identification (RFID) devices (Gubbi et al., 2013). Recent developments in IoT technologies have brought about improvements in consumer products, commercial applications, industrial devices, and other applications for critical infrastructure protection (Khraisat & Alazab, 2021).

In consumer products, IoT devices are used in smart vehicles (networks of moving vehicles), home automation (smart homes), smart cities, wearable devices, and appliances with remote monitoring capabilities (Guhathakurta, 2017). For commercial usage, IoT is applied in medical and health-related applications (Internet of Medical Things, IoMT) for data collection and analysis for research and monitoring (Da Costa et al., 2018; Engineer, Sternberg, & Najafi, 2018). Also known as the Industrial Internet of Things (IIoT), IoT connects industrial devices to acquire and analyze data from connected equipment, operational systems, remote locations, and people equipped with wearable devices. Likewise, in infrastructure, IoT applications monitor and control critical infrastructures like bridges and railway lines (Gubbi et al., 2013). Additionally, the Internet of Military Things (IoMT) applications are deployed in the military domain for national security, reconnaissance, monitoring, and surveillance. Considering its broad scope of applicability, Lu and Xu (2019) predict that the number of IoT devices connected to the internet will reach 25 billion by 2020.

The critical challenge of IoT systems is that they are vulnerable to security threats (Ahmad et al., 2021). The Internet of Things technologies are exposed to severe cybersecurity threats (Singh et al., 2015) and major privacy violations (Howard, 2015). According to Gubbi et al. (2013), the vulnerabilities of IoT networks will significantly increase with complex botnets and denial of service attacks. This problem is compounded because IoT systems are low-powered devices with severe operational limitations and computational power to mitigate malware attacks (Moustafa, Turnbull, & Choo, 2018b). Consequently, malware poses threats to IoT devices’ availability and reliable operation, leading to grave security risks (Nakhodchi, Upadhyay, & Dehghantanha, 2020). Furthermore, a series of attacks target the IoT networks to degrade performance and breach their security using malicious software. Therefore, it is crucial to design robust and accurate methods to mitigate the adverse effect of such attacks (Peters et al., 2020).

Complete Article List

Search this Journal:
Reset
Volume 18: 1 Issue (2024)
Volume 17: 1 Issue (2023)
Volume 16: 1 Issue (2022)
Volume 15: 4 Issues (2021)
Volume 14: 4 Issues (2020)
Volume 13: 4 Issues (2019)
Volume 12: 4 Issues (2018)
Volume 11: 4 Issues (2017)
Volume 10: 4 Issues (2016)
Volume 9: 4 Issues (2015)
Volume 8: 4 Issues (2014)
Volume 7: 4 Issues (2013)
Volume 6: 4 Issues (2012)
Volume 5: 4 Issues (2011)
Volume 4: 4 Issues (2010)
Volume 3: 4 Issues (2009)
Volume 2: 4 Issues (2008)
Volume 1: 4 Issues (2007)
View Complete Journal Contents Listing