A deep learning approach to traffic prediction in emergency situations: the case of COVID-19.

Abstract

Mobile communication has become an important part of the day to day lives that it has influenced greatly every sphere of life starting from the education to even the work place. According to cisco,the number of devices connected to the internet was projected to reach about 28.5 billion in 2022 from 18 billion in 2017. With this in mind, mobile network operators aim at providing a very good quality of service so as to improve the users’ quality of experience with the minimum CAPEX (cap- ital expenditure) and OPEX (operational expenditure). This can only be achieved through network optimization. Network optimization is based on how efficient net- work traffic predictions can be made. Accurate prediction of data traffic in telecom network is a challenging task impor- tant for better network management especially due to the seasonal and emergency situations that cause deviation of the traffic usage from the regular patterns. In this project, we utilized deep learning specifically Long Short-Term Memory al- gorithm to make mobile network traffic predictions. We were able to forecast traffic 32 hours ahead. The model was able to do predictions with the mean absolute error of 0.71 and a root mean squared error of 0.96. These very low values for the errors make LSTMs very suitable for making predictions and adapting to traffic changes.