An Efficient Machine Learning Approach for Detecting Malwares in IoT Environment

In recent research it is observed that malwares in shared system have grown significantly over the previous decade, inflicting significant financial losses to a variety of enterprises. Many anti-virus solution and firewalls have proposed different strategies to guard against these malicious intrusions. Malware's pace, volume, and complexity are creating many new problems to the protection support. The Internet of Things (IoT), a web of millions of objects with sensors that interconnect around wireless or wired stations. The IoT is abruptly expanding throughout the world, but its security is behind. There is a high demand for an effective security architecture capable of protecting IoT systems from potential adversary assaults. Yet, given the dispersed and dynamic disposition of IoT, designing a proper security architecture for IoT is difficult. Machine learning is well adapted to identifying malware on these unsupervised devices, and the findings are encouraging. The machine learning algorithms suffer from excessive computational cost and the potential of privacy leakage. In this paper a machine learning approach for IoT security against malwares is presented. In order to exploit weaknesses in IoT devices, the attackers frequently use aggressive techniques. Grounded on this observation, the recommendations in this paper encourage security solutions of IoT devices to detect and prevent malwares using appropriate ML algorithm(s). This work assesses performance of three implementations namely Bayesian Generalized Linear Model, Boosted Linear Model, and eXtreme Gradient Boosting. Every model generated a malware score based on the improved input characteristics which represents the reliability of an IoT appliance based on a variety of characteristics. OMNeT++ simulator is used to simulate IoT traffic patterns. The acquired findings showcase the efficacy of the recommended plan of action in comparison to other current schemes.