Research And Application Of Web Log Data Based On Association Rule Mining Algorithm

Web Usage Mining is an important research branch in Data Mining area, which researches and analysis the data of web log files to find hidden the association relationship and the users models, so as to improve the Web users personalized service and quality.This paper summarizes the study abroad Web Usage Mining, which details related to concepts and techniques about Data Mining, which further elaborated processing and techniques on Web Usage Mining and Association Rules.In this paper, the target is to analysis of users behavior for improving web site designed, which points out the Apriori algorithm inadequacies for Web Usage Mining. The paper systematic analysis the data characteristics of web log files for topology of web sites, which make web-based topology and compress frequent item sets improvement strategy, the idea of the improvement strategy is to remove the cumbersome home page and relatively cumbersome site page, and apply the classification algorithm CBA of Association Rules to find the affiliated site page through the frequent item sets generated from sub-page, the last frequent item sets generated by adding sit page and home page in sub-page frequent item sets. It saves memory space and reduces the consumption of the system in time. Finally, be accompanied by a comparison example, the detailed analysis to verify the superiority of the new algorithm.