Research And Application Of Web Usage Mining In Web Prefetching

Now Internet has become one of the most important information sources. With the enormous growth of information on the Web, how to get the useful information from this enormous data is becoming an important research topic. Web Mining is the application of Data Mining on the WWW and used to get useful information from the large amount of WWW resources. This information can be used to improve the ervice quality of WWW. So Web Mining has tremendous application values.On the other hand, when surfing on the Internet, users always suffer from long time waiting. Web pre-fetching is one of the most popular strategies which are proposed for reducing the retrieval time by predicting and fetching the most likely web pages that are to be requested soon and improving the service quality of web server (QoS).Through analyzing the Web usage data, Web usage mining can find the patterns how users access the Web. In this process, how to identify the user sessions is one of the most important factors. The statistic based session identification method sets the access interval according to the statistical result dynamicly. The traditional association rules and sequential patterns analysise can't mine the useful rules according to the transactions'hidden classifying characters. But relative supports can solve this problem and be used to Web Usage mining.The Markov model is a history based model which applies the statistical information to Web pre-fetching. This model hasn't considered the web pages'important rank and the the influence of web site structure. The CSM model gets a more precise result by adjusting the transition matrix in the traditional Markov model and improves the service quality of web server (QoS).