Efficient location prediction in mobile cellular networks

Mobile context-aware applications are capable of predicting the context of the user in order to operate pro-actively and provide advanced services. We propose an efficient spatial context classifier and a short-term predictor for the future location of a mobile user in cellular networks. We introduce different variants of the considered location predictor dealing with location (cell) identifiers and directions. Symbolic location classification is treated as a supervised learning problem. We evaluate the prediction efficiency and accuracy of the proposed predictors through synthetic and real-world traces and compare our solution with existing algorithms for location prediction. Our findings are very promising for the location prediction problem and the adoption of proactive context-aware applications and services