Anonymous User Identification Based On Web Browsing Behavior Patterns

With the development of the Internet,Web has become an information network space with massive data resources.Mining web behavior data of users and identifying anonymous web users have important application value.For example,in the field of public security,Web user identification can provide decision support for network early warning,monitoring and so on.Therefore,this thesis intends to explore the identification of anonymous Web users based on their behavior data.(1)User Web behavior pattern mining.Through the analysis of user web behavior data,a user web behavior pattern mining approach based on Bayesian network model is proposed.Experiments show the proposed approach can get the behavior pattern of a specific Web user effectively.(2)User Web behavior pattern clustering.A user behavior pattern clustering approach is proposed and a database of user behavior patterns with classified index is constructed.Experiments show based on the proposed clustering approach and index model,it is effective to determine the user group that a given anonymous user may belong to.(3)Anonymous Web user identification.A Web anonymous user identification approach based on the Bayesian network model's structure,parameters and users time distribution on Web sites is proposed.Experiments show that the approach can effectively determine a specific Web user in a group of given users based on its Web behavior pattern mined.The Web browsing data from Sogou is used to evaluate the proposed approach.Experiments show that the approach is effective to mine behavior patterns of Web users.Moreover,it is efficient to identify anonymous users based on the proposed clustering approach of behavior patterns of Web users.