University of North Carolina
The appearance of Smartphones equipped with various sensors enables pervasive monitoring of mobile users' behaviors and mobility. The Nokia Mobile Data Challenge (MDC) gives one a great opportunity to study the users' mobility models and location pro les from a rich mobile dataset. The realistic data analysis may benefit a wide range of fields from technology innovation to policy making. In this paper, the authors describe their proposed methods to predict the semantic meaning of the "Important places" and the users' next destination based on released MDC data. They explore several features from the sequence of visited places and accelerometer samples, and proposed two types of prediction methods: rule based and machine learning based.