Research Of Web Usage Mining Based On Negative Association Rules

Association rule mining is one of important part in Data mining, Its purpose is to find the connection among the item Sets in a large number of data items. Applying the association rule mining techniques to Web usage mining, we can more effectively organize and administer the site, to find the Page of sites that is interested to users, and recommend to the user real-time, in order to meet the needs of individual users, etc. But At present, most researchers only involves a positive association rule mining in Web usage mining, and less involved in the negative association rules, Negative association rules can find negative associations among the item Sets, it is complement to positive association rule mining techniques. Applying negative association rule mining techniques to Web usage mining, negative correlation Web pages can be found, and may improve the accuracy of Web usage mining. However, how to applying the negative association rules to Web usage mining is a new problem, Therefore, this article discussed the issue, And presents a Web usage mining methods based on positive and negative association rules.This article introduces data mining, Web data mining and Web usage mining, these technologies are discussed, and introduces the concepts and techniques that is related to. The current study and Research Challenges in Web usage mining is focused on Emphasisly, then introduced the positive and negative association rules, we focused on relevant technologies of negative association rules and development status. We Study the related technologies of Web usage mining based on association rules (such as A=>B rules form) and summarize the current study to these. Based on the analysis in the previous。We know that Web usage mining based on positive association rules have more research work but the negative association rules (such as A=>(?)B,(?)A=>B,(?)A=>(?)B rules form) based on Web usage mining is concerned less. Actually, mining the positive and negative association rules at the same time can improve the accuracy of Web usage mining, In this paper, ours experiments can verify accuracy to Web usage mining.so make up the deficiencies in the research that only Positive association rules are refer to.