Machine learning models for wireless network monitoring and analysis

Abstract

The number of smartphones connected to wireless networks and the volume of wireless network traffic generated by such devices have dramatically increased in the last few years, making it more challenging to tackle wireless network monitoring applications. The high-dimensionality of network data provided by current network monitoring systems opens the door to the massive application of Machine Learning (ML) approaches to improve different wireless networking applications. In this paper we evaluate and compare different ML models for the analysis of cellular network traffic, addressing two different and highly relevant problems linked to the end-users and the apps running on their smartphones: detection of anomalies generated by smartphone apps and prediction of Quality of Experience (QoE) for popular apps. We consider an extensive battery of ML models, including single models as well as ML ensembles such as bagging, boosting and stacking. Proposed models are evaluated using real cellular traffic measurements captured at operational networks and at the end devices. Results suggest that decision-tree based models are the most accurate to address these problems, and that collaborative models, in particular stacking ones, are capable to significantly increase performance and robustness of the analysis.