Intelligent Based On The Web Log Mining Site

Along with the development of computer network scale and the inflation of customer need, Internet has become the exchange media by which people publish information, communicate. Although Internet stores a lot of information and provides search engineer, customers can not still find out oneself needing contents.The daily log data of WEB site records lots of visiting path information for the web. We can draw the user hobby information from log file. WEB site's designer proceeded the page reorganization and even intelligent web can predict the next visited page in the future.The paper introduces the classification of data mining technique first and analysis the system construction of web log mining .The paper emphases preparing-process. Secondly we try to introduce the notion of association rule into the web mining system and provide similar Apriori.In this paper, it brings forward five methods about abstracting rules from the frequency item. We select the best method through the experiment. In the practicality, the paper provides two matching rule idea of choosing appropriate rules.In the last, we propose recursive Cluster algorithm to find some URL Sets. The remanding engineer predicts the next URL from the fit URL sets.