Predicting smartphone app usage with recurrent neural networks

Abstract

Nowadays millions of apps are available and most of users install a lot of apps on their smartphones. It will cause some troubles in finding the specific apps promptly. By predicting the next app to be used in a short term and launching them as shortcuts can make the smartphone system more efficient and user-friendly. In this paper, we formulate the app usage prediction problem as a multi-label classification problem and propose a prediction model based on Long Short-term Memory (LSTM), which is an extension of the recurrent neural network (RNN). The proposed model explores the temporal-sequence dependency and contextual information as features for prediction. Extensive experiments based on real collected dataset show that the proposed model achieves better performance compared to the conventional approaches.