Research on DoS Attack Traffic Detection for Industrial Control System

The rapid development of the Internet has brought convenience to the industrial control system, but also brought serious security problems of network attacks. DoS attack is the most common network traffic attack. The traditional abnormal traffic detection method has the problems of long model training time and high false alarm rate. Aiming at these problems, this paper proposes an optimization algorithm based on decision tree CART, as well as feature coding and feature classification of data. The experimental results show that the decision tree model has better performance than other machine learning models, and achieves a higher accuracy of 98.2 %.