Research On Traffic Volume Prediction Of Improved ARMA Under Wavelet Transform

After entering the 21 st century, with the development of science and technology and network, telecommunications industry has developed fast in our country, especially the three big operators—China Mobile, China Unicom and China Telecom. They keep expanding the network size and enriching types of business to increase the customer quantity significantly. In this industry, traffic volume is an important concept, it’s size not only concerns the size of the customer’s communication quality, but also provides a basis for the development of operators. Although these factors including the population of the ground base stations, the performance of the network equipment and the call times of users, directly determine the size of the traffic volume, what we have to do is how to accurately predict in advance according to the known history data of traffic volume, in order to make adjustments in time and avoid the potential risks.History data of traffic volume is a kind of time series, we can use a lot of mature time series models to forecast, but how to predict more accurately is an important research topic. In this article the idea of wavelet transform is introduced into the ARMA model, we get a new improved ARMA model and then predict. Based on this, main research contents of this article are as follows:① Firstly we introduce the ideas of wavelet transform and single branch reconstruction algorithm in detail.② Then we lay special stress on several prediction models of traffic volume,including AR model, MA model, ARMA model.③ The original sequence consisting of traffic volume data which has collected is decomposed into approximate series and several detail series by wavelet transform.Every series is reconstructed into new sequence by single branch reconstruction algorithm. These new sequences are predicted by respective ARMA models and then to estimate the original sequence.④ Based on the theoretical analysis about the forecasting model, this article illustrates the framework of the ARMA prediction model. The result of ARMA prediction model is compared with the result of ARMA prediction model with the wavelet transform.⑤ Design idea of prediction system of traffic volume in telecommunications industry is systematically put forward.What has been mentioned above is not only theoretical guidance to predict thetraffic volume for communication operators, but also beneficial beginning of the engineering application on prediction of traffic volume in telecommunications industry.