A Network Traffic Anomaly Detection Method Based on Deep Learning

The scale of network traffic is becoming increasingly large and the attack methods are constantly evolving. Traditional network traffic anomaly detection methods are facing challenges. So, this paper proposes a network traffic anomaly detection method based on Attention Mechanism and Spatial Convolutional Network with Bidirectional Long Short Term Memory (AMSCN-BiLSTM). This method integrates spatial convolutional network (SCN) to extract spatial features from network traffic. To enhance feature capture capability, an attention mechanism is introduced and combined with bidirectional long short-term memory network (BiLSTM) to encode sequence data, effectively capturing long-term dependencies. Comparative experiments verify that the network traffic anomaly detection method based on AMSCN-BiLSTM can quickly identify abnormal behaviors in network traffic. This effectively provides a solution for network security protection.