A trajectory prediction approach for mobile objects by combining semantic features

In this paper, we propose a trajectory prediction approach for mobile objects by combining semantic features. Firstly, the geographic trajectories of all users are transformed to the semantic behaviors trajectories. Then the semantic trajectory pattern sets are extracted. The common behavior of mobile users is analyzed in semantic trajectories and the users are clustered based on the semantic behavior similarity, by which geographic trajectory pattern sets are discovered. Based on the semantic trajectory pattern sets of individual users and the geographic trajectory pattern sets of similar users, the STP-Tree and SLP-Tree are constructed. By indexing and partly matching on the two pattern trees and introducing a weigh function, our method can predict a user's recent move position. The proposed method can effectively extract users' behaviors and adjust inaccurate prediction results compared with the methods using only geographic features. Experimental results on a large number of real-world and synthetic data sets show that the precision of our method are significantly improved compared with the state-of-the-art methods.