Assessment of Traffic Prediction Models for Mobile Communication Networks

Abstract

In mobile communication systems, radio access network cells transport data in resource elements. Each cell has a maximum capacity of resources per second, depending on its bandwidth. Resource elements time series express the variations of traffic in a cell, which varies strongly, both spatially and temporally. In this work, the prediction of mobile communications traffic, based on available monitoring data, is addressed. Were used two strategies to predict the future observations of the time series: Holt-Winters and SARIMA.