Neutrosophic Set and Machine Learning Model for Identifying  Botnet Attacks on IoT Effectively

Wasal S AL-Bash AL-Azzawi; Hassan W. Hilou; Nawfal H. warush; Hasan Meslmani; Ahmed  A El-Douh; Ahmed Abdelhafeez

Authors

Wasal S AL-Bash AL-Azzawi Computer Engineering Technology, Al-Salam University College, Baghdad, Iraq
Hassan W. Hilou Computer Engineering Techniques, Al-Ma'moon University College, Baghdad, Iraq
Nawfal H. warush Mechanical Engineering, Al-Nahrain University, Baghdad, Iraq
Hasan Meslmani Medical Instrumentation Techniques Engineering Department, College of Engineering Technology, Ashur University, Baghdad Iraq
Ahmed A El-Douh Applied Science Research Center, Applied Science Private University, Amman; Information Systems Department, Faculty of Information Systems and Computer Science, October 6 University, Giza, 12585, Egypt
Ahmed Abdelhafeez Applied Science Research Center, Applied Science Private University, Amman; University, Giza, 12585, Egypt 7Computer Science Department, Faculty of Information Systems and Computer Science, October 6 University, Giza, 12585, Egypt

Abstract

Botnet attacks, in which attackers utilize reciprocal communications between IoT
devices to undertake extensive harmful actions, are one of the most significant risks in WSNs. In
this sense, advancements in the realm of dependable and effective defenses against this kind of
threat—specifically, trustworthy techniques for detecting, recognizing, and thwarting botnet
attacks—are becoming more and more significant and pertinent. This work offers a thorough
analysis that successfully detects botnet assaults on the Internet of Things by using machine
learning techniques, including Random Forest and LSTM. These algorithms are examined,
contrasted, and demonstrated to be very successful in identifying intricate patterns suggestive of
botnet activity, leading to a notable enhancement in IoT security. The goal of the study is to help
solve the issue of WSN and IoT security in general. The neutrosophic set is used in this study to
overcome uncertainty information. We triangular neutrosophic model to select the best model.
The results show RF is the best compared to other models.

DOI: 10.5281/zenodo.15717644