Abstract: User trajectory behavior analysis based on mobile phone signaling data has been widely used in many business fields such as traffic monitoring, urban management, and population analysis. Aiming at the shortcomings of mobile phone base station positioning deviation and lack of semantic information, this paper proposes a positioning method based on geographic information. Firstly,we construct the specification of geographic entity data, then establish the weight mapping relationship between geographic entities and base stations by clustering the mobile phone signaling data, and fit the user trajectory to the actual geographic entity using spatio-temporal reasoning and deep learning technology, and finally realize the method.The verification results show that this method improves the accuracy of user positioning and provides reliable data support for the application of subsequent business scenarios.

Key words: mobile phone signaling data, trajectory restoration, user positioning, geographic entities