Adjustments based on wavelet transform ARIMA model for network traffic prediction

Time series can be decomposed into different spectrum sub-sequence using wavelet decomposition, and restoring all the time series prediction from the decomposition can effectively improve the prediction accuracy. This paper presents a method of network traffic prediction named AWARIMA. First we choose a 5-layer db3 wavelet to decompose network traffic data,make the similar sub-sequences from the wavelet decomposition to the merger and then apply the appropriate length of the historical data sequence in different frequency bands to meet the requirements of ARIMA models. Results show that predictive effect of AWARIMA method is markedly improved compared with traditional time-series models.