Web-based Data Mining Website Optimization Design Applied Research

In recent years,with the rapid growth of Internet in flow,scale and complexity,Web has become a huge,widely distributed and global information service center. Web brings a wealthy information and great facilitation,some urgent problems that need to be solved appeared,one of them is personalized information services.One of the ways to directly or indirectly solve this problem is that using Web log mining techniques is used in Web personalized service. Web log data is effectively mined, which can preferablely help us to find frequent user accessing paths ,which is very importment to provide users with personalized service.The information diversity in the www makes it even harder for users to find the desired information. How to find potential and interesting knowledge from enormous data is a very important and meaningful issue.Web log mining is that using the data being produced when users are communicating with servers to find connotative and disciplinarian knowledge by data mining technology.We can acquire the frequency and behavior model when users visit the site.Using the frequency and behavior,we can obtain the user's preferences and accessing habits, optimize the Web site structure and the hyperlink structure between the Web pages,improve the service quality of site,and ameliorate the site performance and provide users with better services.Through the behavior analysis of users'visit, we can recommend the visiting path with the shortest time. First, based on the personalized website model, comprehensively considering the visiting time and saved times of users, we can provide a measuring calculating method based on users'preferences. Secondly, we present a sociation rules mining method WARMVI based on vetor inner, which could help us find websites with high relations. Considering users'own interests, experiment simulations give out the recommending stategies by means of the analysis of users'visiting modles so as to inhance the visiting efficiency. The innovations of our work are mainly five aspects as follows:(1)deeply research on the definition , basic principle , methods and category of web log mining , and describe the application of web log mining detailedly. (2)analyse the characteristics and difficulties of web log mining, and investigate the individuation of web site and browsing mode for individual user and mass users.(3)describe the the process of web log mining, and analyse the four phases for preprocess: data cleanup, user recognition,dialog recognition and transaction recognition, and also present relevant improved algorithms.(4)estimate the preference of users by means of the time for browsing and the times for accessing, on the basis present a new Web Association Rules Mining based on Vector-Inner, by which reduce the time for searching the correlative web pages.(5)design a experimental system consisted of web server, access database and web log to validate the validity of association rules algorithm.The in-depth study on behavior of browsing web site, the analysis of Web site performance, and the improvement of structure between web topology and page hyperlinks make the better service become possible. Web Log Mining is an application of data mining in web server log to obtain the pattem and the aeeess behaviorial mode of the users.This helps to improve web sites trueture,its aeeess quality and its performances.