Mitigating Iot Botnets With CNN-LSTM And Anomaly Detection

Abstract

The rapid expansion in use IoT has made extreme security threats with botnet attacks leveraging device vulnerabilities to carry out malicious actions like DDoS, data theft, and network interference. Traditional intrusion detection systems (IDS) fail to keep up with the pace of threat growth. This paper is a full survey of deep learning-based detection of IoT botnets and presents a hybrid approach offering detection efficiency and accuracy. Through the use of CNN, LSTM, RNN, and ensemble methods, the approach scans host and network traffic to offer a scalable adaptive solution. Experimental outcomes on benchmark datasets offer superior performance compared to the traditional IDS in terms of accuracy, reduction in false positives, and efficiency of computation. Real-time deployment and self-adaptation to new threats are left as future work.