| With the rapid development of mobile Internet,online services have become an essential part in people's daily life.People usually have several accounts on multiple and diverse online services such as Weibo,QQ and Taobao.Each account is a digital identity of the user.Meanwhile,the data generated by users is rapidly increasing.Therefore,the most intriguing question is how to leverage this big data for a better and deeper understanding of each individual user.However,in the vast amount of data,information of a user is always fragmented.To make user behavior analysis and user profiling become more complete,consistent and continuous,it is effective to link up all the traffic data of the same user in different online services.This thesis aims to propose a model to link user-'s all digital identities.Existing methods are usually designed for a specific service domain or at best service domains that are semantically similar.However,in order to obtain an integrated profile for each individual,it is necessary to link his/her digital identities of different services across multiple service domains together.In contrast,our goal is to address the most general case in which data across service domains is separately generated and has obvious differences in characteristics.To addiress the problem,this thesis proposes a digital identity linkage model.It derives several significant attributes from users'online behaviors,such as various fingerprints of terminals,spatio-temporal behavior of users,and leverages a supervised classification method to discover the relationship between users' different identities.By using real-world network traffic collected from a large province in China,we evaluate the model and the linkage precision and recall all achieve 99%.Especially,the inputs of model,i.e.,network traffic flows,cover all online behavior of users who connect with Internet through monitored networks,which makes it possible to link digital identities of users in whole online world. |
PDF Full Text Request