Mobile User Environment Detection using Deep Learning based Multi-Output Classification

Abstract

Future mobile networks can hugely benefit from cognition of mobile user behavior. Indeed, knowing what/when/where/how the user consumes their mobile services can notably improve the self-adaptation and self-optimization capabilities of these networks and, in turn, ensure user satisfaction. The cognition of mobile user behavior will thus help 5G networks to face the variable consuming habits of users which in turn impact the network conditions, by predicting them in advance. In this paper, we focus on the "where" part, i.e., the detection of the environment where a given user consumes different mobile applications. A statistical study on the real activity of users reveals that there are multiple various environment types corresponding to the mobile phone usage. A Deep Learning based model is introduced to intelligently detect the user environment, using supervised and semi-supervised multi-output classification. Relevant multi-class schemes are proposed to efficiently regroup the multiple environment categories in more than two classes. We empirically evaluate the effectiveness of the proposed model using new real-time radio data, gathered massively from multiple typical and diversified environments of mobile users.