Network ddos attack detection method based on artificial intelligence

Abstract

The present invention discloses a network DDoS attack detection method based on artificial intelligence, including: obtaining an attack traffic anomaly dataset by performing abnormal fluctuation screening and peak comparison on network traffic data; performing source address geographic location mapping based on the dataset to determine a geographic distribution range of abnormal traffic; analyzing inter-regional connection relationships to construct a connection topology and judging and extracting a topology structure corresponding to a coordinated action pattern; extracting dynamic change patterns and temporal correlation based on the topology to obtain decentralized collaboration characteristics; performing cluster analysis on source address behavior patterns based on these characteristics to determine a control pattern recognition result; if centralized control is recognized, performing deep tracing on communication path convergence points to obtain attack intent inference basis; constructing an overall profile of attack behavior by integrating the basis and distribution features, and generating a DDoS attack detection result. The present invention achieves accurate inference of complex DDoS attack strategy intents, thereby improving detection accuracy and early warning capability.