Performance Evaluation of Botnet Attack Detection Using XAI
Usha Divakarla, K. Chandrasekaran · 2024
Botnets have become a major security concern in recent years, and there is a growing need for more advanced detection systems that can accurately and effectively identify botnet traffic. In this research paper, we propose a novel XAI botnet detection system that is built on top of an Artificial neural network (ANN) architecture and leverages several XAI techniques to improve model interpretability and help identify specific features that contribute to botnet traffic detection. We demonstrate the effectiveness of our XAI botnet detection system by training it on the new CIC DDoS 2019 dataset. We compare the performance of our XAI botnet detection system against several traditional machine learning models and demonstrate that our system outperforms them in terms of both accuracy and interpretability. One of the key advantages of our XAI botnet detection system is its transparency and interpretability. In conclusion, our research paper demonstrates the potential of XAI for improving botnet detection accuracy and interpretability. We show that our XAI botnet detection system outperforms traditional AI models in terms of accuracy and interpretability, and is more transparent and human-centric. As AI applications continue to evolve and become more complex, XAI models will become increasingly important for ensuring the safety and security of our digital infrastructure