Recurrent Neural Ne tworks for Predicting Mobile Device State

JUAN MANUEL RODRIGUEZ; ALEJANDRO ZUNINO; ANTONELA TOMMASEL; CRISTIAN MATEOS

Encyclopedia of Information Science and Technology, Fourth Edition

IGI Global

Lugar: Hersey, PA; Año: 2017;

Predicting mobile devices states is a crucial issue in different domains, such as context-aware applications, computational offloading, or ad-hoc networks, among others. The current state of a mobile device can be regarded as the consequence of the previous sequence of states. Interestingly, Recurrent Neural Networks have shown an enormous potential for predicting sequences. In this context, this work evaluates the suitability of such neural networks for predicting the future states of mobiles devices. Particularly, the approach focuses on predicting battery level, charger connection state and WiFi network connection. Experimental evaluations conducted on data gathered from real smartphone usage have demonstrated the adequacy of neural networks for performing highly accurate predictions.