Information security threats can occur across the network and can provide relatively faster threat detection and prevention. This article examines the likelihood of threats by collecting and analyzing network traffic data. This article discusses the integration of intelligent technologies for processing and analyzing network flow data. A model is proposed that combines logistic regression to improve anomaly detection and prediction accuracy. Functioning tables are used to formalize network processes. Experimental results are presented for both normal and anomalous network flow cases, demonstrating that the use of network flow information is effective for network security.