Based On Web Log Mining Of Association Rules In E-government Applications

Web log records the users' information who visits the web site, it contains massive routing information. To analysis them is beneficial to grasp the users' fondness and visit habit, and may be useful to the website structure's optimization and reorganization. Analyzing log data follows several ways, for example, we can count data form, statistic pages collection being often visited, count the important table data often needs analyzed, and find out the same visit route by analyzing website .Web log mining is the Analysis and processing to the server log using the data mining thought, so, it can solve each kind of problems proposed above.Firstly, This paper introduces the web data mining and web log mining's specific contents as well as theirs membership; Meanwhile, it has carried on the analysis and the research on the web log mining's pretreatment technology, described each duty of the traditional data pretreatment stages and designed relative data tables and pretreatment flow algorithm. Secondly, the paper describes the data mining's function and algorithms common used, studies FP-growth algorithm with emphasis in the association rule, and compares the performance merits between FP-growth and Apriority algorithm. After that, the paper has carried on the analysis to the classics FP-growth algorithm, proposed a new improved algorithm. Through the experiment contrast, it points out that the improved algorithm has saved the time with higher performance. Thirdly, based on the original ahead e-government systems, the paper discusses the application of the improved mining algorithm in the e-government web log mining system. Through research and improvement on the association rule algorithm, the paper uses the improved FP-growth algorithm to carry on the data mining to the click stream data pretreated, and analyzes the mining results. Accordingly, giving advice to the website structure construction and the government system's improvement.In the summary and forecast parts, it discusses the issues which need to be further perfected in the present research work, and points out the direction of future research.