Wireless Traffic Prediction Based on Bayesian Seasonal Adjustment Algorithm

Faced with the rapid growth of network traffic in the development of wireless network. To guarantee a better user experience in wireless network, the accurate and timely estimation of traffic is becoming more important. It is necessary to expand the network capabilities of 4G cells reasonably according to the future traffic growth trend of the cell. However, some cells face the traffic demand of small-scale users, which is more easily disturbed by seasonal changes and irregular factors than by a traditional traffic forecast. In this paper, we use the additive model of Bayesian seasonal adjustment algorithm to decompose the historical monthly traffic into trend, season, and irregular components. Finally, we use three models to predict the components and sum the predicted results to get the monthly traffic prediction results. The results show that the proposed method is reasonable and effective. The prediction results have been applied to the decision-making of expansion of 4G network.