Description:
The growing number of connected Internet of Things (IoT) devices has led to the daily growth of network botnet attacks. The networks of compromised devices controlled by a single entity can be used for malicious purposes such as denial of service distributed IoT botnet attacks and theft of personal information. The weak security measures of many IoT devices make them easy targets for compromise and inclusion in botnets. In this research, we propose a system for detecting botnet attacks. We develop an ensemble learning system to detect botnets in network traffic with high-performance scores. The system will analyze the traffic and identify any suspicious behavior that may indicate the presence of a botnet. For this purpose, we use the benchmark CTU-13 dataset to build the applied machine learning and deep learning techniques for comparison. We propose a novel ensemble technique, K-neighbors, Decision tree, and Random forest (KDR), to achieve high performance for botnet attack detection. Study results show that the proposed KDR gives 99.7% accuracy in 12.99 s. Hyperparameter optimization and k-fold cross-validation are employed to substantiate the performance. Our research study contributes to the body of knowledge on the detection of botnet attacks and provides a practical solution for securing IoT devices against botnet attacks.