I'm Only Unhappy when it Rains: Forecasting Mobile QoS with Weather Conditions

Abstract

Global increase in the use of mobile Internet service generates interest in mobile network studies to determine and forecast the QoS provided by mobile operators. This study proposes different methods to forecast signal strength, one of the most important mobile Internet QoS indicator, based on time series analysis and considering external information about weather conditions as temperature, humidity and precipitations due to the effect they cause on mobile Internet QoS. This work shows the feasibility of forecasting mobile signal strength using crowd data corresponding to mobile devices in Santiago, Chile and that the inclusion of weather information generates more accurate forecast models for a given geographic area, obtaining good performance by all models used at comparing their forecast error values for weekly predictions. To the best of the authors' knowledge this is the first attempt of using weather information together with real data gathered from user devices in order to forecast mobile signal strength.