Research And Implementation Of Personalized Page-display Optimization Technology In Mobile Terminals

After rapidly evolving traditional web searching technology was widely accepted by users, Mobile Internet search with its greater convenience becomes the heat of attention. Retrieving information with mobile terminals is the future development trend.Storage space,network bandwidth , screen size, page processing speed, and battery power of mobile termination affect the user experience. Now, the network bandwidth is being improved, the restrictions of storage capacity, page processing speed and battery life will be solved along with the development of hardware technology. Only the screen size which directly influences on effects mostly among all is difficult to change .Thereby improving terminal displays of the Web page is imperative. In view of the present insufficiency in mobile terminal page display, this paper designs personalized page-display optimization system.Firstly, this paper introduces the related theory of mobile terminal page displays for the farther research, and then describes the current relevant technology about mobile terminal page display: Web page structure extraction, Web page content extraction, and page layout mode in mobile terminals, they are prepared for the design of page-display optimization system.Secondly, according to the personalized page display strategy, designs work flow of the system and the overall framework. Proposing Page-displaying algorithm based on RBF block importance degree learning and Text extraction algorithm based on the predicted contents. In view of the preceding algorithm,the paper uses RBF neural network algorithm And since people with different personalities have different needs, the paper adds users'interesting factors to block importance degree model to determine what displays on the screen, it shortens users'position time. In view of the latter algorithm, in consideration of integrity of semantic content and basing on the idea of the tag ratio extraction algorithm, it can predicts"jump"which can identify the page textby the dynamic change trend of the tag ratio to. And then based on distributing features of the content, the paper puts forward SK-Means clustering algorithm which can effectively separate useful information from noisy content.Finally, deploying the system to Windows Mobile5.0 simulator, and evaluating the system's performance tovalidate the system's performance.This paper is an important part of "a new version of the network terminal operating system research and development community" of Ministry of Science and Technology. It has an important significance both in theory and reality.