PSTN Traffic Model Based on Wavelet Decomposition

In this paper,PSTN traffic mathematic model—Poisson distribution,is reviewed. Poisson model does not adapt to analysis of the new pattern of PSTN data.Wavelet transform is very important tool to forecast and to model PSTN data.Through making full use of the multiresolution analysis in the wavelet transform,the PSTN datatraffic is divided into different bands of frequency by using wavelet decomposition,which can simplify the frequencies of the decomposed traffic and smooth the traffic.Then Regressive Model is employed to predict the sub-traffic. Experimental prediction result with the real PSTN data traffic indicates that the model improves old model's accuracy and provides a relatively accurate prediction.