MTP-NT: a mobile traffic predictor enhanced by neighboring and transportation data

Abstract

The development of techniques able to forecast the mobile network traffic in a city can feed data driven applications, as VNF orchestrators, optimizing the resource allocation and increasing the capacity of mobile networks. Despite the fact that several studies have addressed this problem, many did not consider neither the traffic relationship among city regions nor information from public transport stations, which may provide useful information to better anticipate the network traffic. In this dissertation, we propose a new deep learning architecture to forecast the network traffic using representation learning and recurrent neural networks. The framework, named MTP-NT, has two major components: the first responsible to learn from the time series of the region to be predicted, and the second one learning from the time series of both neighboring regions and public transportation stations. The work also reviews the 5G infrastructure based on open 3GPP specifications to explore ways to implement the framework in a real architecture. Several experiments were conducted over a dataset from the city of Milan, as well as comparisons against widely adopted and state-of-the-art techniques. The results shown in this work demonstrate that the usage of public transport information contribute to improve the forecasts in central areas of the city, as well as in regions with aperiodic demands, such as tourist regions. Thus, this research seeks to evaluate the performance of traffic forecasting models using public data, in order to validate the performance gain with the aggregation of public transport data. The aggregation of unconventional data can be a way of adding information to the model through input that has not been explored in the scope of this research area.