WIMAX traffic forecasting based on neural networks in wavelet domain

Abstract

In this article we propose an approach for predicting traffic time series based on the association of the Stationary Wavelet Transform (SWT) with Artificial Neural Networks (ANN). We focused on comparing the quality of forecasting obtained using different configurations of the ANN. We tested our different configurations using real traffic data recorded at each base station that belongs to a WiMAX Network developed by Alcatel. We compared our approach with previously forecasting models using ANNs and showed the performance of our neural network configuration.