Research On Anonymization Based Privacy Preserving Method On Geosocial Networks

With the development of Internet and big data technology,more and more location information and social information are collected,stored and analyzed by research institutions and enterprises through geosocial network under the background of artificial intelligence and intelligent city.In order to make better use of the potential value of data,data holders will publish the collected data by data sharing for researchers to mine and analyze.However,data records from geosocial networks contain highly sensitive personal information,so privacy issues become a bottleneck restricting the publication of geosocial network data.At present,anonymity technology is a common method in the research of privacy preserving data publishing,which hides the quasi-identifier of the individual in the data to avoid the attacker getting sensitive information of the target as far as possible.Both in privacy protection of geographic location and social networks,many research works have migrated and extended anonymization methods.However,geosocial network data is heterogeneous data with complex data connections.Its complexity makes the previous anonymization methods unable to effectively protect users' privacy,resulting in serious threats to privacy security.Therefore,for the anonymization framework of geosocial network data,location and social information need to be processed simultaneously to protect individual sensitive information.Our main contributions are as follows:Firstly,most current geographic location anonymity methods which only consider location points and do not consider the relationship between locations,exist the problem of losing location patterns.To solve this problem,we formally define a frequent location relational graph model based on top m locations model,propose a privacy protection scheme for anonymization of geographic location data based on the contact graph,and design two anonymization algorithms.In the process of anonymity,the scheme considers the feature of co-occurrence patterns between geographic locations,which ensures that the anonymous geographic location data sets can maintain pattern characteristics between locations,and improves the usability of data mining and analysis.Finally,we verify the effectiveness of the scheme on two real data sets from the aspects of privacy protection performance and data usability.Secondly,because geosocial networks are rich in location and social data,anonymous protection only for location can't resist the mixed attack of location and social information.Therefore,on the basis of the first work,we introduce social information and design two anonymization algorithms based on graph modification strategy,which protect the privacy of users in geosocial networks,reduce the anonymity loss and improve data usability.Finally,different performance indicators are evaluated and analyzed through a series of experiments,and the efficiency of the algorithm is verified.