The Method Of Mining Inter-transaction Association Rule For Web Usage Mining

Web data mining is the process of extracting valuable information in the Web environment according to the user's browsing behavior. Web usage mining is an important component of data mining and the user is the core of Web usage mining. Web usage mining extracts the meaningful and descriptive models of user behavior by the algorithm of association rules, enhances the efficiency of searching information and provides users with instructional forecast experience.By analyzing and summarizing the previous mining algorithm of inter-transaction association rules, two new inter-transaction association rules are proposed:(1) Mining algorithm of inter-transaction association rules based on clustering analysis. According the algorithm, initial and complex data sets are reduced by clustering analysis, firstly, which avoid multi-scans of database and plentiful false rules generation. This step generates many small data sets which are treated by inter-transaction association rules. Experimental results show the method forecasts users'interest more efficient and accurate than those methods which only use inter-transaction association rules.(2) Mining algorithm based on clustering analysis'inter-transaction association rules and dual-strategy interest-analysis mode. The proposed algorithm used the dual-strategy analysis model to judge the integrity of inter-transaction association rules and divided initial database into Association database and Markov database, which avoided the false rules. After that, the clustering analysis is adopted to remove plenty of redundant data at Association database and Markov database, which improves algorithm's efficiency. Finally, the prediction of association rule and Markov model are respectively added to each sub-database. Experimental results show this algorithm is superior to the traditional business association rules inaccuracy and efficiency.With the development of the Web technology, the dates in the Web-server are going more and more huge and complex. How to solve the contradiction between efficiency and accuracy of data mining is a new challenge. Thus, this paper uses the inter-transaction association rules to improve accuracy of data mining, and employs the technology of cluster analysis to improve efficiency, which provides new idea for Web usage mining.