A study on cellular wireless traffic modeling and prediction using Elman Neural Networks

Modern cellular resource management for dynamic control of channel resources and energy efficiency improvement relies largely on early and accurate monitoring and prediction of cellular base station traffic volumes. Analysis of network traffic volume over space and time plays an important role in traffic prediction. In this paper, we examine both the temporal and spatial characteristics of cellular traffic data generated by users in a large population city in China. We analyze and cluster base-stations of similar characteristics. We determine the sliding window sizes and integrate the Elman Neural Network (ENN) after applying wavelet transform in order to realize traffic volume prediction. We present numerical results to illustrate the accuracy of wireless traffic volume prediction, and we test the performance of our method to demonstrate improvement over some existing methods.