Research On Instant Messaging User Geolocation Technology Based On Cross-platform Association

Instant messaging(Instant Messaging)is a class of applications that can send and receive Internet messages instantly.After more than 20 years of development,it has developed into a comprehensive information platform integrating communication,information,entertainment,search,and e-commerce.The instant messaging user geolocating technology can achieve the positioning of the target user without being detected by the target,and is of great significance in the practical application of specific target tracking,public opinion spread analysis,and punishment of illegal crimes.Due to different user location protection strategies adopted by different instant messaging tools,existing instant messaging user geolocating technologies often design corresponding geolocating algorithms for a specific platform,and there is a problem of a limited number of users geolocating.Therefore,this paper proposes an instant messaging user location technology based on cross platform Association,which overcomes the limited number of users based on a single platform by comprehensively using the user attribute information and user trajectory information of multiple instant messaging tools,and proposes two cross platform user association location algorithms and a user information acquisition method.The main work of this paper is as follows::1.The existing geolocating methods based on the relationship between the system reported distance and the actual distance of the user have a limited number of geolocating users.Therefore,this paper proposes an associated positioning method based on matching user attribute information.,this paper proposes an association and geolocation method based on user feature information matching.First,by analyzing the internal structure of the instant messaging tool,based on the user interface element module to accurately find user name,gender,age and other characteristic information,the problem of automatic and accurate acquisition of user information in the multi-source instant messaging system is solved.Secondly,based on the mobile phone number search and friend recommendation function,the user information obtained through the mobile phone number is associated with the crawler information.Finally,based on the obtained system notification distance information,a high-precision geolocation algorithm is used to geolocate users.A 30-day geolocating experiment was conducted for We Chat,Momo and QQ users in Zhengzhou,China and New York,USA.The results show that: with the existing typical geolocating algorithm based on space partition and geolocating algorithm based on advanced trilateration in comparison,the number of user positioning has increased by 31.0% and 14.2% respectively.2.The existing matching algorithms based on user trajectory often use the frequency information of the geographic location where users appear to match users across platforms,making it difficult to accurately characterize the sequence of geographic locations in the trajectory.Therefore,this paper proposes an association and geolocation of user based on trajectory spatiotemporal feature matching.Firstly,obtaining the user location information based on the user geolocating algorithm,and constructing the user track spatiotemporal domain point set according to the user's occurrence time.Secondly,the spatiotemporal domain points in the track are divided according to a certain time granularity and distance scale,and the user track is represented by a grid sequence.Finally,the user track is transformed into a vector based on the TF-IWF weight calculation model.A matching experiment was conducted on 12,102 user trajectories collected from 515 pairs of We Chat and Momo users with associated relationships.The results show that: compared with the existing typical K-BCT(k-Best Connected Trajectories)user trajectory matching algorithm and user trajectory matching algorithm based on TF-IDF(Term FrequencyInverse Document Frequency)model,the user trajectory vector obtained by the proposed method can more accurately express the spatiotemporal characteristics of the user trajectory,and the accuracy and recall of trajectory matching can reach 94.6% and 96.9%.3.The existing methods of user information acquisition based on optical character recognition may make mistakes in information extraction.Therefore,this paper proposes the GUI elements searching based information automatic acquisition.Firstly,analyzing the program structure of the instant messaging tool to obtain the Activity information of the "nearby" interface.Secondly,automatically searching and injecting the information module in the user graphical interface to obtain the content of the module such as user name,announcement distance,signature etc.and then save the information as editable text information.Finally,executing script program to obtain information automatically,at the same time,the user feature information of the specified area is acquired through the simulation of location.The user information acquisition experiments were conducted for We Chat,QQ and Momo in Zhengzhou and New York.The results show that the accuracy of information acquisition of the proposed method can reach 96.1%,which is 29.2% higher than the user information acquisition method based on optical character recognition,and the efficiency has also improved.Finally,we conclude this thesis,and point out the problems that need to be further studied.