Discovering user's trends and routines from location based social networks

Abstract

ABSTRACT: Location data is a powerful source of information to discover user's trends and routines. A suitable identification of the user context can be exploited to provide automatically services adapted to the user preferences. In this paper, we define a Dynamic Bayesian Network model and propose a method that processes location annotated data in order to train the model. Finally, our model enables us to predict future location contexts from the user patterns. A case study evaluates the proposal using real-world data of a location-based social network.