DEVELOPMENT OF AN INTELLIGENT BOTNET DETECTION AND MITIGATION FRAMEWORK USING MACHINE LEARNING TECHNIQUES

The rapid evolution of cyber threats, particularly botnet attacks, poses significant risks to network security, challenging traditional defence mechanisms. This study presents the development of an intelligent botnet detection and mitigation framework using machine learning techniques. A dataset comprising multiple malware families and botnet captures was collected, pre-processed, and used to train three classifiers such as Decision Tree (DT), Artificial Neural Network (ANN), and Support Vector Machine (SVM). These models were integrated into an ensemble classifier to leverage their complementary strengths for enhanced detection accuracy. Performance evaluation revealed that the SVM achieved the highest individual accuracy of 0.97, while the ensemble model demonstrated superior detection capability, attaining an accuracy and recall of 0.99. System integration using a simulated wireless network showed that the ensemble classifier effectively identified and isolated malicious nodes, outperforming traditional firewall solutions. The study concludes that ensemble-based machine learning approaches provide a robust and reliable solution for botnet detection, offering a scalable framework for enhancing cybersecurity in heterogeneous network environments.