Personalization Services Based On The WEB Mining

As the rapidly growing of the network technology, The information in the Internet goes up by index, It's difficult to find the necessary one in so much information, Even if we can find them for they are sometime mix up with noise, How to find out the user's necessary information rapidly is a hot study for recent learners. The WEB personalization is just result from this requirement. The WEB personalization means a WEB site dynamically prescribes the browsing information, and provides browsing suggestion for users according to their preference ,and the direct way is to recommend and lead the users. Recently there're many kinds of personalization services systems, Such as Association Rules, Collaborative Filtering and Content Filtering based on their practical technology ,This essay mainly puts analysis on the study of Association Rules and Collaborative Filtering in the personalization services.This thesis firstly introduce the concept and definitions of Data Mining ,WEB Mining , Personalization services and so on .In the light of the recent approach to count the user's interests, This thesis analysis its pros and cons; Puts forward another method with combination of explicit and implicit approaches; adds some factors which haven't been taken into account in the common methods, thus puts forward the relevant counting formulas.Secondly this thesis introduces the data preprocessing ,Association Rules mining; Analyses some shortages exist in the Association Rules mining; Puts forward the WEB page Association Rules mining based on the user's interest ,which combine the user's interest with the WEB page Association Rules mining; and applies it to the personalization service, and proves that this method would improve the recommendation precision through many experiences.Then it also introduces the definition and the counting methods of Collaborative Filtering algorithm. It also gives introduction of the methods of dealing with the Collaborative Filtering by use of interest ,thus collects the users with similar interest to form the similar user's Clustering ,therefore ,puts the personalization services into practice and illustrates the application of Collaborative Filtering by some practical examples.Finally in the light of the close and sparsesty problem in the Collaborative Filtering ,this thesis also puts forward the Clustering analysis method to form similar interest user's groups to solve this problem and it also provides the K-means Clustering counting methods based on the interesting .