Predicting the next location: A self‐attention and recurrent neural network model with temporal context

Abstract

Abstract Nowadays, the popularity of mobile devices and location‐based services have generated a large amount of geographic data. It provides the opportunity for researchers to employ techniques to predict the next location. However, predicting the next location is difficult because it depends on temporal and spatial factors, and it is closely related to the historical behavior of users. In this article, we first analyze the geographic data of users and discover the potential behavior patterns of users. Then, we mine the relationship between user's movement behavior and temporal feature. Hence, we propose a method based on a recurrent neural network and self‐attention mechanism to predict the next location where users may visit. Our model can explore sequence regularity and extract temporal features according to historical trajectories information. Experimental results on a real‐world dataset demonstrate the effectiveness of our proposed model.