Classification of IOT-Malware using Machine Learning
Sanjay Madan, Monika Singh · 2021 International Conference on Technological Advancements and Innovations (ICTAI) · 2021
Every day, attackers target embedded IoT devices, causing damage to key cyber-infrastructure, obtaining users’ personal information, and misusing it to a greater extent. Data confidentiality, authentication, and privacy, denial of service, nonrepudiation, and digital content protection are only a few of the difficult security challenges that must be handled. To exploit these resource-constrained Cyber-physical systems, attackers use brute-force assaults, man-in-the-middle attacks, injecting malicious code, eavesdropping, and backdoors, among other methods. In this research, we present a hybrid analysis method for analyzing Linux-based IoT malware and event correlation for incident management using anomaly detection. For malware classification, the machine learning model is built using information from both static and dynamic analysis of harmful programs. For anomaly identification and event correlation, the anomalous DDoS traffic detection approach is also proposed. The F1-score is maximized for various DDoS attacks using the threshold selection approach, and the results are compared to the state-of-the-art literature.