A feature analysis approach to network traffic in communication networks

Network traffic holds the highlighting dynamic features. How to accurately characterize the hidden properties of network traffic has an important impact on network activities, such as anomaly detection and performance analysis. This paper proposes a feature analysis approach to describe network traffic. Firstly, the wavelet packet transformation is used to extract the multi-scale feature of network traffic. Then the principal component analysis method is exploited to refine the hidden features of network traffic in the time-frequency domain. Finally, to validate our feature analysis method, an anomaly detection test is conducted. Simulation results show that our approach is promising.