A novel mobility prediction scheme for outdoor crowded scenario using Fuzzy C-means

Forecasting the users movement and behavior is extremely valuable for communication networks to support the explosive mobile data in the outdoor crowded area, in the respects such as network deployment, resource allocation and mobility management. Due to the large users number and complex individual behavior, it is difficult to accurately predict the user's movement. In this paper, we take an academic campus as a study example and propose a novel mobility prediction scheme based on data mining algorithm. First, we divide the whole area into several prediction areas based on the number of mobile users, and divide the prediction time into several periods according to the scenario feature. Then, we classify the trajectories of the mobile users into groups based on Fuzzy C-means (FCM) clustering, and discover the frequent mobility patterns in each prediction area at different periods using sequence pattern mining. Finally, we determine the group for the new user and find the most matched mobility pattern to predict its future location. Simulation results show that the proposed scheme achieves a better performance compared with exiting schemes in terms of the handoff numbers and dwell time.