HYBRID MACHINE LEARNING MODEL FOR EFFICIENT BOTNET ATTACK DETECTION IN IOT ENVIRONMENT

Abstract

The rapid proliferation of Internet of Things (IoT) devices has significantly increased the attack surface for large-scale botnet intrusions, demanding intelligent and adaptive detection mechanisms. This study presents a novel LSTM-driven botnet detection framework designed to model the sequential and temporal characteristics of network traffic in IoT environments. Unlike traditional signature-based or standalone deep learning approaches, the proposed mechanism leverages the memory-gated architecture of Long Short-Term Memory networks to effectively capture long-range dependencies and evolving attack behaviors within network flows. The model is evaluated using the UNSW-NB15 dataset, which comprises diverse attack categories and normal traffic patterns. Comprehensive experimental analysis demonstrates that the LSTM-based mechanism achieves superior classification accuracy, improved precision–recall balance, and reduced false alarm rates compared to conventional ANN and CNN-based classifiers. The framework autonomously extracts meaningful temporal features without manual engineering, enhancing adaptability to previously unseen botnet variants. Furthermore, its lightweight architecture supports scalability and near real-time deployment in resource-constrained IoT infrastructures. The results confirm that the proposed mechanism strengthens detection robustness, improves generalization capability, and provides an efficient defense solution against dynamic and sophisticated botnet threats in modern IoT networks.