Enhancing Cyberattack Detection in IoT Environments Through Advanced Resampling Techniques

Abstract

As the world increasingly relies on emerging technologies like the Internet of Things, there is a growing demand for large-scale distributed software to perform various tasks, facilitate communication, and share resources between devices. However, the implementation and configuration of such softwares can create openings for intrusion attacks through vulnerabilities and weaknesses. To address this concern, we have developed a machine-learning solution that leverages Logistic Regression and Random Forest classifiers with data balancing techniques to classify intrusion attacks accurately. Our experiments demonstrated the most effective results using the Random Forest classifier and oversampling techniques.