Prefetching web pages for improving user access latency using integrated Web Usage Mining

The tremendous growth of web has resulted in overloaded servers and network congestion. Millions of users access the internet simultaneously and limited bandwidth causes a bottleneck to improved service. Fast response time is essential to keep the users at bay and thus techniques need to be improvised to reduce the latency of accessing web pages. Web caching, in this regard, offers an effective solution by temporarily storing files nearby, thereby reducing the access time, but limits the system to an extent. A more powerful solution, web prefetching, has been devised, that fetches pages in advance, further reducing the access latency. Numerous solutions to web prefetching have been proposed in the past, integrating various web and data mining techniques. In this paper, the techniques of Web Usage Mining have been integrated for analyzing the proxy server logs, to get an insight into the user access patterns and generate rules. Link analysis method has been proposed to be integrated with these rules in order to rank the pages to be prefetched.