Semantic web access prediction using WordNet Lenka Hapalova (supervised by Ivan Jelinek) hapall11@fel.cvut.cz (jelinek@fel.cvut.cz), Czech Technical University in Prague, Faculty of Electrical Engineering - Dpt. of Computer Science and Engineering, Karlovo nám. 13, 121 35 Prague 2, CZ Abstract. The user observed latency of retrieving Web documents is one of limiting factors while using the Internet as an information data source. Prefetching became important technique to reduce the average Web access latency. Existing prefetching methods are based predominantly on URL graphs. They use the graphical nature of HTTP links to determine the possible paths through a hypertext system. Although the URL graph-based approaches are effective in the prefetching of frequently accessed documents, few of them can pre-fetch those URLs that are rarely visited. In our paper we aim to propose a new prefetching algorithm that would increase the efficiency of Web prefetching and that will embody the new demands for Web personalisation and Web search assistance. The aim of the research is to design a system for web page prefetching. The system should use user’s link path history in combination with the semantic path history. To enable this, semantically annotated web pages are necessary. We cannot rely on the web documents’ creators thus one part of the work must be the design and implementation of simple annotator based on WordNet just for purposes of our research. Keywords: Web access latency, prefetching, semantic Web, Web access prediction, personalisation, Markov models 1 Introduction Due to the rapid development of the Internet usage and the exponential growth of online information, the Internet has become one of the most important information sources. The usage of World Wide Web (WWW) as a data source has increased as it provides quick and easy access to a tremendous variety of information in remote locations. The wide range of sources’ locations is the benefit as well as the drawback of the WWW. Users often suffer from long delay time when they access Web pages – so-called Web access latency. With the rapid growth of Web services on the Internet, users are experiencing access delays more and more often. Document pre-fetching is an effective tool to improve the access to the World Wide Web. In comparison with caching, pre-fetching aims to pre-retrieve Web documents (more generally Web objects) to the client side even before they are