期刊名称：International Journal on Computer Science and Engineering
印刷版ISSN：2229-5631
电子版ISSN：0975-3397
出版年度：2012
卷号：4
期号：08
页码：1420-1426
出版社：Engg Journals Publications
摘要：In this study we uses an approach for increasing web performance by analyzing and predicting the user behavior from user access log , identifying the missing paths in the logs and completing the same using a web structure so as to enhance the accuracy of web usage mining. This paper uses a simulated example to compare the result of the Pre-fetched pages after completing the missing paths in the logs using web structure and if this paths are not completed. The result shows the accuracy of the predicted and pre-fetched pages has improved after completing the missing paths in the logs. This paper uses Markov Model 1st order and 2nd order to predict most frequently accessed pages.
关键词：Server log; User session behavior; Markov models and Association rules.
