Abstract: Aiming at the problem that the traditional model could not accurately identify and detect network traffic anomalies, we proposed a network traffic anomaly detection model based on time series analysis. Firstly, the original data of network traffic was extracted, and the original data was denoised by wavelet threshold to eliminate the influence of interference factors. Secondly, time series analysis method was used to mine the relationship among network traffic data, and network traffic anomaly detection model was established. Finally, simulation experiments were used to verify the effectiveness and superiority of the detection model. The result shows that time series analysis can accurately and timely detect abnormal behavior of network traffic, and the detection results are better than other current network traffic anomaly detection models.

Key words: traffic anomaly, network security, echo state flow, time correlation, detection model