Research And Implementation Of Analysis System Of Network User Virtual Identity

With the rapid development of the Internet,the style of human life is undergoing earth shaking changes,and the way to communication is diverse.Accordingly,a new concept—virtual identity comes into being.Different from real life the Internet community,which is composed of virtual identity,is a virtual world that cannot be seen or touched.People in different countries and different colors can cross the barriers of geography and language,and they can communicate freely and easily on the Internet.In the network,the communication is carried out by digital symbols and various virtual expressions,and it is realized through the indirect dialogue between human and computers.In the virtual world people can change their name,gender,age and background freely to cover up their real informations,so it is a hot pursuit of people.Various applications and sites in the Internet are required user to apply for a unique user account to distinguish different users,and the most users are using false information to registration,some users will register multiple user accouts with different false informations,therefore,there are massive virtual identities in the Internet.For this reason,we will analyze which network virtual identities belong to the same person to analyze the user behaviors.We adopt the MapReduce programming model and Hadoop Distributed File System on Hadoop to analyze due to the huge data.Firstly,this thesis introduces the basic concept of virtual identity and the status of network user behavior analysis.Secondly,this thesis introduces the foundation platform of this research,including HDFS,MapReduce and HBase.Thirdly,a detailed description of the framework of analysis system of network virtual identity,including three modules:data acquisition module,data preprocessing module and data analysis module.Finally,the implementation of the system on the Hadoop platform is described in detail.