Predicting Location Trajectories of Humans by Their Diverse Social Ties

Abstract

The location prediction issue impacts a wide range of practical areas ranging from urban planning to epidemic controlling. Numerous previous studies regarding location prediction have been carried out based on both individual historical trajectories and traces of the dense social ties of individuals such as friend and in-role. However, a kind of the so-called Familiar Stranger social tie has been discovered and identified recently, which was neglected previously and may improve the location prediction. In this paper, we propose a novel location prediction method which first takes the trajectories of the familiar stranger of individuals into account. We validate our method to achieve better performance with multiple social ties to predict locations of users with three empirical human traces datasets.