Network Traffic Prediction Based on Wavelet Transform and Extreme Learning Machine

Aiming at the defects of the current network traffic prediction model, a novel network traffic prediction model based on wavelet transform and extreme learning machine is proposed. At first, wavelet transform is used to process time series of network traffic to get different frequency components, and each feature is used to be reconstructed by phase space, and secondly extreme learning machine is used to model and predict network traffic of each component, and predicted value of each component are reconstructed by traffic by wavelet transform to get network traffic prediction results, at last, the performance is tested by network traffic data and compared with other models The results show that the proposed model can get higher accurate prediction results than other models, and the prediction results of network traffic are more stable, it provides modeling tool for network traffic.