User's Location Prediction in Location-based Social Networks

The wealth of user-generated data in Location-Based Social Networks (LBSNs) has opened new opportunities for researchers to model and understand human mobile behaviour, including predicting where they are most likely to check-in next. In this paper, we propose a model that leverages the use of Global Temporal Preferences and Spatial Correlation, to help make predictions for a previously unseen user - the so-called cold-start problem. The experimental results on a real-world LBSN dataset show that our proposed model outperforms the state-of-the-art approaches on prediction accuracy and can alleviate the cold-start problem.