The Analysis Of User Behaviors On Internet Website Access

With the increasing popularity of the Internet, the Internet economy is in a competitive status, the demand of websites’ service quality becomes higher and higher and the management of network service providers is also from blind to science. Thus, the analysis of user behaviors on website access has become increasingly important. By analyzing the Internet users’ behavior, we can grasp the regularity characteristics of users’ access to websites and apply the access patterns into websites. It has very important significance on clustering, classification, etc.The precondition of analyzing Internet user behaviors is to identify users clicking.In recently years, the network advertisement has become more and more popular. The browser will automatically generate a large number of ads requests when users have a link request. This has caused great disturbance for our analysis. Therefore, how to effectively remove these ads interference has become an important research topic. At the same time, web surfing is the most important behavior of network behaviors. We also need research network user access behavior based on accurately identifying users clicking behavior.Firstly, this thesis proposed a new method to analyze the advertising networks by constructing link-tree of web sites. We examined large number of popular web sites using this link-tree based method to reveal the actual status of internet advertising in China, from which we found two kinds of online advertising architectures:two-tier and three-tier architecture. We also analyzed the traffic characteristics of internet advertisement component, which shows some differences among different categories of web sites. Secondly, there exist many methods to identify users clicking behavior which have a lot of shortcomings. We proposed a new cutting algorithm which is based on the difference of different kinds of nodes. After executing the cutting algorithm, it will form the real user access tree.Lastly, web surfing is the most important network behavior,so this thesis analyzed the characteristics of network behaviors on enterprise network users.We analyzed the average depth,the average access times of homepage in an access tree respectively in different fields. At the same time, the top10websites which has the largest clicking mount have also been analyzed.