Neural nets based predictive prefetching to tolerate WWW latency

With the explosive growth of WWW applications on the Internet, users are experiencing access delays more often than ever. Recent studies showed that prefetching could alleviate the WWW latency to a larger extent than caching. Existing prefetching methods are mostly based on URL graphs. They use the graphical nature of hypertext links to determine the possible paths through a hypertext system. While they have been demonstrated effective in prefetching of documents that are often accessed, they are incapable of pre-retrieving documents whose URLs had never been accessed. We propose a context-specific prefetching technique to overcome the limitation. It relies on keywords in anchor texts of URLs to characterize user access patterns and on neural networks over the keyword set to predict future requests. It features a self-learning capability and good adaptivity to the change of user surfing interest. The technique was implemented in a SmartNewsReader system and cross-examined in a daily browsing of MSNBC and CNN news sites. The experimental results showed an achievement of approximately 60% hit ratio due to prefetching. Of the prefetched documents, less than 30% was undesired.