Based Web User Interest And Demand For Technology

With the exponential growth of information in the internet, to get the accurate and necessary information from internet for users is becoming more and more urgent. User interesting mining as an efficient technology to discover potential and valuable information from vast web information is indeed emerging and under the spotlight, to some extent, it resolved the contradiction between a variety of internet information and the focus of user's demands. The main study content is how to accurately dig individual consumer's interest, build the model and analyze the interest based on the users'browsed information or behaviors or web log etc. Further, we mine the interest of user groups through clustering the users based on individual interest, and describe users'interest needs from individual interests and group interests to perfect the interests'model. This paper specifically analyzes and introduces the following proposed technologies and algorithms through the project of user interests mining based on Web and WAP, their effectiveness is checked by real data's experiment.(1) The technology to build user interests modelThe text analysis based on content is the basis of mining users'interests in this paper; we crawl and parse the main body of webpage based on the URL of users'browsed history. In the phase of getting text and preprocessor time, we proposed the method of extracting main page based on rules and DOM tree, and the method of classify the texts based on chi-square and the weight of key words; In the phase of user interests modeling, we proposed the diversification of modeling ideas, not only build model through long-term interests and short-time interests, but also synthesize the modeling thinking of separately statistic their own history behaviors and some users'similar behaviors from individual interests and group interests, this way describes users'interests all-sided.(2) The technology of mining users'individual interestsWe get users'interest tendentious in information demands mainly through mining the users'history interviewed WebPages. This study proposes two algorithms suit for the application scenarios in this topic: ①the algorithm of mining users'long-term interests②the algorithm of mining users'short-term interestsUsers'interests reflected in the preference of theme demands and content form, in the process of maintenance and modification of the users'interest, using the method of interest's category interview density plus time attribute to mining the long-term or short-term interest. In the aspect of updating user's model, we introduce the forgetting factor, using biological forgotten law to forget for the sleepy users or interests, strengthen the memory of active users or interests, finally reach the purpose of updating users'interests'models.(3) Mining group users'interestsThere exists similar action between some users, this kind group interests can be regards as the individual user's potential interest, so can say it has the function of recommendation. The difficulty of group interests mining is the efficient clustering for vast users.