Application Research Of Data Mining In Web Personalized Service

Challenged by the mass information on WWW, how to extract knowledge quickly and correctly that people interested in, tacit, potentially useful from the web data of dynamic, heterogeneous and half-structure has become a topic of general concern.Web personalized service is that websites could examine users' log on information by tracking users' log on, and then find users' preferance in order to provide dynamically browsing suggestions and contents for users. Web data mining is that finds the potential and useful model or information in web data by using data mining technology.In this paper, fuzzy cluster analysis is used in web mining. Once the data of log is getting processed, we carry on web transactions data of web fuzzy cluster, analysis, studying, then extract effectively the model in which users are interested. By means of using web mining technology, the organization and display of the information that is shown in the website could be improved according to users' interests and preferances. Web mining can also provide different severing strategies and contents for different users, improve the quality of websites. And then the website construction and modification could be done better. The times of invalid click would be reduced. The time spent on searching information would be shortened and the visiting effiency would be improved, too. Finally web personalized service would be better realized.This paper first studied the process of web data preprocessing, and the correlation algorithm of all assignments has been given out.By the way of researching the fuzzy cluster's theory and technology, we have proposed the fuzzy cluster's model of web usage mining. We carry on exploratory research of application and realizing in such aspects as web user's cluster and web page cluster in web transactions at this model. Studied the method of finding frequent browsing path based on page level and preference.Synthesizing the XML technology, a web personalization system module has been put forward. The correlation manage system based on the module can adjust web service tactics and offer pertinence service according to the interest to users, so as to make the website more personalized with no increasing in existence service website's hardware equipment.