Towards Spatio-Temporal Aware Real Location Restoration for Signaling Data

This study introduces the novel problem of Real Location Restoration, aimed at reconstructing accurate user trajectories from the coarse-grained mobile signaling data. To tackle this problem, we propose the Spatio-Temporal Aware framework for Signaling Data (STASD), a pioneering approach that encodes the complex spatiotemporal relationships of cellular trajectories. Leveraging a unique global transition graph, STASD captures high-order spatial relationships to effectively mitigate the Ping Pong Effect, a common issue in signaling data analysis. Our extensive experiments showcase the framework's capability to accurately restore real-world trajectories, significantly advancing the field of mobile data analysis by providing a novel method to interpret and utilize signaling data for detailed location insights.