Research On The Application Of Data Mining Technology In Personalized Web

With the development of internet, the visitors frequently lose their course in a complex information space because of the interference caused by a large number of Web links. However, the visitors desire that the website should recommend personalized information and they could accurately and quickly obtain the useful information from the large number of web links. Data mining technology can provide a strong support for personalized web service.This paper deeply analyzes the principle, strengths and weaknesses of a variety of personalized recommendation technology and conclude the trend of personalization technology development.Firstly, considering non-structural and non-semantic of network information, This paper presents a personalized information services recommendation model based on semantic, and the model can provide a personalized recommendation platform for better personalized service through seeking the information matched with the current user personalized information. Secondly, using data mining techniques to the personalization systems, and this paper proposes a association rules discovering algorithm based on intersection maximum frequent patterns. The algorithm can find large item sets quickly and avoid a large number of candidate sets. Finally, after mining the frequent user mode by association rule discovering algorithm, and this paper uses cluster analysis to classify the user mode and mine users access patterns with similar interests, for excluding the non-relevant user-mode. Then more accurate user interest patterns can be generated to provide recommendations for the current user. This application that association and cluster analysis method are used for personalized web provides a theoretical basis for the personalized recommendation services. The experimental results show that:the clustering algorithm based on association is effective and can improve the quality of personalized web services.