User's Interests Based On The Video Stream Mining Model Design And Simulation Implementation

As the Internet has been used widely in every field of life and the electronic commerce developed fleetly, web advertising has become the main battlefield of e-commerce. However, traditional web advertising was so extensive that it not only leaded to the waste of advertising fee but also disturbed the users’normal browse on the web page because it couldn’t gain the their interests. So, getting what are the web users interested in and doing accurate Ad service are hot spots of the Internet.On the other hand, as the network bandwidth rises fast, the network speed has been increased at substantially and the transport quality is guaranteed, and so the technique of video streaming has been evolving fast and been applied widely. Therefore lots of applications based on the video streaming came out. However, as the leaders of telecommunications field, the mobile network operators always acted around the edges in this field. For this reason, they are trying lots of new ways to change their profit model which is based on fee charging for bandwidth, of which directed Ad service has gone to the popularization step stage.Users’interests digging is the premise and foundation of directed Ad service. Although mining the users’interests directed at the web text has been very mature, the related technology based on the video streaming is in exploration. This thesis introduces the technology of getting text and users’behavior information, and lists some representation of users’interests and personalized recommendation techs and systems.Then it discusses what information can be got and how to get it in router environment. With the information, we design a special interest model for video field, which is named Standard Category Tree. The SCT model is a kind of improved ontology model. To employ this method, the first step is building a standard category of the video field. In every category, we store some information, for example, keywords list of this category and related attribute probability of the audience, which is used to calculate the audiences’goods interests.It shows how to design the model of SCT(Standard Category Tree) to present the audiences’interests. To verify the validation and veracity, it designs a demo system, which contains building the standard classification tree of the video category, putting the user’s watching records to right category, calculating the probability of user’s crowd kinds and estimating how is the user interested at the goods.At the end, this thesis analyses and concludes the advantage and weakness of SCT, and expects the future of users’interests mining based on video streaming.