Effective Detection of Cybersecurity Attacks in IoT Networks using Random Forest Classifier Integrated with mRMR Feature Selection
Sami Ur Rahman, Department of Computer Science, Islamia College Peshawar, Pakistan.
Atif Khan, Department of Computer Science, Islamia College Peshawar, Pakistan.
Corresponding Author:
Atif Khan(atifkhan@icp.edu.pk)
Abstract:
This research examines the cybersecurity concerns in IoT networks using the Bot-IoT dataset, the Random Forest classifier, and the mRMR feature selection technique. This study focuses solely on the Bot-IoT dataset, which is notorious for its class imbalance and lack of samples, as opposed to earlier research that utilized obsolete datasets. In spite of these obstacles, the Random Forest classifier and mRMR feature selection technique are shown to be successful for increasing the detection rate of intrusion detection systems (IDS) for the Bot-IoT dataset. To undertake a comprehensive examination of the IDS system, the study utilizes both binary and multi-class categorization. Our results demonstrate the importance of utilizing the Bot-IoT dataset and the efficacy of the Random Forest classifier and mRMR feature selection method for evaluating cybersecurity assaults in IoT networks.
Keywords:
Bot-IoT; mRMR; Random Forest; Intrusion Detection