一种基于模式匹配度的用户移动规则挖掘及位置预测方法研究

he location prediction of mobile user is effective for the temporal and spatial allocation of system resources, and it can improve the resource utilization of the mobile communication system. The traditional location prediction method has a problem of low prediction accuracy due to unreasonable calculation of support. To overcome these challenges, this paper presented a method called mining user mobile rule based on pattern matching degree and location prediction. It can predict location with the base coverage mesh as the unit for mobile users in mobile communication systems. The detailed procedure contained three phases: a) mining user mobile patterns from graph traversal; b) generating user mobile rules from user mobile patterns; c) predicting based on user mobile rules. Finally, the conducted experiments on 10 batches of trajectory data demonstrate that this method has the characteristics of fewer rules, high support, and high confidence than traditional methods, and highly predict accuracy.