Users Behaviour And Intention Analysing Based On Internet Traffic Data

After years of rapid development,the number of Internet users in China has a large scale exaggeration,China's Internet environment is increasingly mature,and its application is also constantly changing from single to wide.At the same time,with the development of the Internet,its impact on people is also more depth and wide,the mark people made in Internet is also more and more rich.Under this background,we are deepening in the era of big datastep by step,forming a vast amount of data of Internet users.With the expansion of massive network data,the previous data processing methods have been unable to meet all the use scenarios.However,Hadoop platform with distributed processing frameworkprovides us a very good way,and becomes an important tool for the processing of massive Internet data.The thesis first introduces the main content of the research,its purpose and significance.From the term of influence and meaning of Internet behavior,it describes Internet user behavior and intention.Then it introduces briefly the Hadoop platform and its programming framework.After that,this thesis focuses on the snalysis model in the research,deep belief networks,support vector machine and term frequency-inverse document frequency algorithm for text processing.And it describes the training methods of these models.Then this thesis focuses on the analysis of user behavior and intention and describes our accurate prediction models.The resuls provides support for user behavior and intention snalysis.At last,it summarizes the advantages and problems of the analysis and put forward the future development of the analysis.