Abstract: World Wide Web is an important area for data mining research due to the huge amount of information. The success of the WWW depends on response time. Predictive prefetching is an important technique to reduce latency. To predict the user request, millions of web logs from client side or from server side or from proxy side need to be analyzed. Most of the existing predictive prefetching methods are based on URL dependency graph, Keyword search techniques and Link structure without considering all the factors for cleaning web logs. In this study, we propose a new predictive prefetching framework based on our new preprocessing algorithm for cleaning web logs and user session identification. Our analysis shows that the new cleaning algorithm reduces latency considerably than the cleaning techniques used by the existing predictive prefetching methods. The experimental results based on five different servers prove the importance of cleaning of web logs before using them for predictive prefetching process.
G. Arumugam and S. Suguna , 2008. Predictive Prefetching Framework Based on New Preprocessing Algorithms Towards Latency Reduction. Asian Journal of Information Technology, 7: 87-99.