User Behavior Analysis On Mobile Internet

With the development of science technology, mankind has stepped intoage of mobile Internet. Accessing into Internet with mobile terminals hasbecome a new way to surfing in nowadays. In Japan and western world whosemobile Internet is developed. In China, the mobile Internet is developing,however, the large user group shows that the mobile Internet has been apotential market and immeasurable value for research on the field aboutmobile Internet, and the research on it is very significant for the developmentof mobile Internet evolution in China and the benefit of enterprise in theInternet.This paper from real user data, proposes a mobile Internet user behaviorresearch method that could be composed by three parts: basic network indexanalysis, service traffic model analysis and web-user relation model analysis.Under this method, the user behavior on mobile Internet is studied, andaccording to the user data this method is proved.In the part of basic network index analysis, according to the scenario ofthe telecom carrier, firstly a data capture mechanism on mobile Internet isgiven, and the data captured by this mechanism is used to analyze user traffic,user clicks, online time. According to the three index study, the property of"burst" and "heterogeneous" in mobile Internet user behavior is found,which is compared with the property on traditional Internet.According to the index study, the basic user behavior phenomenon isclear, however it is still not be represented by math. In the part of servicetraffic model analysis, based on the mobile Internet transport mechanism, theindexes of mobile Internet web page traffic are selected to used to build themath model that could be matched by classic math model, therefore, there isa mathematic model could well interpret what happens on mobile Internet.For the further study on the influence of user behavior on mobile Internetwebsite and service, in the part of web-user relation model analysis, thispaper gives a Web-user model that is composed by user and web dimension.In the user dimension, the mobile Internet user access model and user serviceparameters are showed. Firstly, steady determiner is brought to give the regulation that the user access to the mobile Internet. After that, K-meanscluster algorithm is used to pick out the user data that is positive in thenetwork, in order to count the value of user click, user access and online timeon different services. In the website dimension, a new method is introducedto indicate website development and competition state in mobile Internet,and this method is proved by the user data, at the same time, the developmentand competition of websites in mobile Internet is observed.Through the study on mobile Internet user behavior, the method on userbehavior study is proposed, which could be the reference for the future study.Under this method, the properties of "burst" and "heterogeneous" in mobileInternet user behavior are found. After that, based on these properties themobile Internet webpage traffic model is built, which could help the trafficmodel study in the future. At last, through the Web-user model, the regulationof user accessing to mobile Internet is studied, the value of attribute ondifferent network service is given and the method that analyze webdevelopment and competition in the mobile Internet is showed.