IoT-POT: Machine Learning-based Detection of Mirai Botnet Attacks in IoT
Anshika Sharma, Himanshi Babbar · 2024
The Mirai botnet attack has become a serious threat to Internet of Things (IoT) devices because it can undermine network security by launching large-scale attacks by taking advantage of weaknesses. This paper uses the IoT-POT dataset, a vast collection of network traffic data collected by IoT honeypots, to suggest a machine learning (ML)-based method for detecting Mirai botnet infections. By utilizing supervised learning methods such as K-Nearest Neighbour (KNN), Logistics Regression (LR), Decision Tree (DT) and Random Forest (RF), create strong models that can recognize the harmful behaviour of Mirai botnet attacks. The IoT-POT dataset facilitates the creation of efficient detection algorithms by offering insightful information about the characteristics and conduct of Mirai attacks. Evaluate the models’ ability to correctly discriminate between legitimate and malicious IoT communications through a thorough testing and evaluation process. The results show how well ML works to identify Mirai botnet attacks and emphasize the value of preventative security strategies in securing IoT ecosystems from new online threats.